While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

diddly-squat said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

I want to know how the car knows when the light is red or green, and if it is “looking” how it knows where there light stanchion appropriate for it is located.

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

Driverless cars read traffic lights including turning arrows.

dv said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

Driverless cars read traffic lights including turning arrows.

How is this done. And particularly how do they find the correct lights? There are many places in city areas where there are multiple lights rather close together.

buffy said:

I want to know how the car knows when the light is red or green

I think that would be literally the simplest part about designing a driverless car

dv said:

buffy said:

I want to know how the car knows when the light is red or green

I think that would be literally the simplest part about designing a driverless car

I don’t. Masses and masses of visual confusion at a fiveways intersection.

buffy said:

diddly-squat said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

I want to know how the car knows when the light is red or green, and if it is “looking” how it knows where there light stanchion appropriate for it is located.

driverless cars have video capability, they can ‘see’ what is around them

diddly-squat said:

buffy said:

diddly-squat said:

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

I want to know how the car knows when the light is red or green, and if it is “looking” how it knows where there light stanchion appropriate for it is located.

driverless cars have video capability, they can ‘see’ what is around them

Unless it’s a white truck on an overcast day…

poikilotherm said:

diddly-squat said:

buffy said:

I want to know how the car knows when the light is red or green, and if it is “looking” how it knows where there light stanchion appropriate for it is located.

driverless cars have video capability, they can ‘see’ what is around them

Unless it’s a white truck on an overcast day…

yeah, well it seems they are no worse than people in this regard…

diddly-squat said:

poikilotherm said:

diddly-squat said:

driverless cars have video capability, they can ‘see’ what is around them

Unless it’s a white truck on an overcast day…

yeah, well it seems they are no worse than people in this regard…

Still safer in fact.

poikilotherm said:

diddly-squat said:

poikilotherm said:

Unless it’s a white truck on an overcast day…

yeah, well it seems they are no worse than people in this regard…

Still safer in fact.

indeed… bring on the driverless cars I say…

diddly-squat said:

poikilotherm said:

diddly-squat said:

yeah, well it seems they are no worse than people in this regard…

Still safer in fact.

indeed… bring on the driverless cars I say…

They could actually replace couriers for that matter and instead of leaving a you weren’t home note in the letterbox, phone your contact number with an automated voice telling you they are out the front

Cymek said:

diddly-squat said:

poikilotherm said:

Still safer in fact.

indeed… bring on the driverless cars I say…

They could actually replace couriers for that matter and instead of leaving a you weren’t home note in the letterbox, phone your contact number with an automated voice telling you they are out the front

I think that within two decades driverless cars will essentially replace all man driven vehicular transport… worldwide we’re talking millions of jobs here

diddly-squat said:

Cymek said:

diddly-squat said:

indeed… bring on the driverless cars I say…

They could actually replace couriers for that matter and instead of leaving a you weren’t home note in the letterbox, phone your contact number with an automated voice telling you they are out the front

I think that within two decades driverless cars will essentially replace all man driven vehicular transport… worldwide we’re talking millions of jobs here

You could probably set it up so the drive-less car is used constantly throughout the day ferrying people all over the place, it could even belong to a specific suburban block and costs shared.

I don’t think I’ve actually got an answer yet. So far people have just said they can do it. How does a driverless car locate and read traffic lights, particularly at intersections where they are not directly across (ie, not 90/180 degree crosses) and where there are arrows?

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

How do Google’s self-driving cars work?

Google’s driverless car has eight sensors,

The most noticeable is the rotating roof-top Lidar – a camera that uses an array of 32 or 64 lasers to measure the distance to objects to build up a 3D map at a range of 200m, letting the car “see” hazards.

The car also sports another set of “eyes”, a standard camera that points through the windscreen. This also looks for nearby hazards – such as pedestrians, cyclists and other motorists – and reads road signs and detects traffic lights.

Speaking of other motorists, bumper-mounted radar, which is already used in intelligent cruise control, keeps track of vehicles in front of and behind the car.

Externally, the car has a rear-mounted aerial that receives geolocation information from GPS satellites, and an ultrasonic sensor on one of the rear wheels that monitors the car’s movements.

Internally, the car has altimeters, gyroscopes and a tachometer (a rev counter) to give finer measurements on the car’s position. These combine to give the car the highly accurate data needed to operate safely.

The biggest worry would probably be protecting them from being hacked either remotely or injecting malicious code into their system when being constructed

Another good article

How Google’s Self-Driving Car Works

buffy said:

I don’t think I’ve actually got an answer yet. So far people have just said they can do it. How does a driverless car locate and read traffic lights, particularly at intersections where they are not directly across (ie, not 90/180 degree crosses) and where there are arrows?

Driverless cars use a combination of video cameras and imaging arrays to detect and read things about their environment. This includes signage and, of course, traffic lights.

Cymek said:

The biggest worry would probably be protecting them from being hacked either remotely or injecting malicious code into their system when being constructed

Or the morality built into the programming:

http://www.scientificamerican.com/article/driverless-cars-will-face-moral-dilemmas/

I read this a while ago. It wasn’t something I’d actually thought about before. I understand what they are saying about the car doesn’t have morals and doesn’t “decide” as such. But the programmer does, back in the development.

buffy said:

Cymek said:

The biggest worry would probably be protecting them from being hacked either remotely or injecting malicious code into their system when being constructed

Or the morality built into the programming:

http://www.scientificamerican.com/article/driverless-cars-will-face-moral-dilemmas/

I read this a while ago. It wasn’t something I’d actually thought about before. I understand what they are saying about the car doesn’t have morals and doesn’t “decide” as such. But the programmer does, back in the development.

this being said, driverless cars are in a far better position to make the best decision when it comes to collision avoidance. They have far more information as to the environment, know more accurately how the car will perform or how quickly it will come to a stop and can do all of this in a fraction of the time it take a person to make the same decision.

CrazyNeutrino said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

How do Google’s self-driving cars work?

Google’s driverless car has eight sensors,

The most noticeable is the rotating roof-top Lidar – a camera that uses an array of 32 or 64 lasers to measure the distance to objects to build up a 3D map at a range of 200m, letting the car “see” hazards.

The car also sports another set of “eyes”, a standard camera that points through the windscreen. This also looks for nearby hazards – such as pedestrians, cyclists and other motorists – and reads road signs and detects traffic lights.

Speaking of other motorists, bumper-mounted radar, which is already used in intelligent cruise control, keeps track of vehicles in front of and behind the car.

Externally, the car has a rear-mounted aerial that receives geolocation information from GPS satellites, and an ultrasonic sensor on one of the rear wheels that monitors the car’s movements.

Internally, the car has altimeters, gyroscopes and a tachometer (a rev counter) to give finer measurements on the car’s position. These combine to give the car the highly accurate data needed to operate safely.

In addition to all the sensors, laser and video, maps are also used

buffy said:

Cymek said:

The biggest worry would probably be protecting them from being hacked either remotely or injecting malicious code into their system when being constructed

Or the morality built into the programming:

http://www.scientificamerican.com/article/driverless-cars-will-face-moral-dilemmas/

I read this a while ago. It wasn’t something I’d actually thought about before. I understand what they are saying about the car doesn’t have morals and doesn’t “decide” as such. But the programmer does, back in the development.

Hopefully the removal of the human aspect reduced fatalities quite significantly and the moral dilemma the car may face doesn’t happen very often, I suppose the needs of the many outweigh the needs of the one. I wonder how often a human has to make the decision to hit one person instead of many and they aren’t impaired in some way.

Thank you diddly, that is the sort of information I was looking for. From the second link:

>> Two things seem particularly interesting about Google’s approach. First, it relies on very detailed maps of the roads and terrain, something that Urmson said is essential to determine accurately where the car is. Using GPS-based techniques alone, he said, the location could be off by several meters.

The second thing is that, before sending the self-driving car on a road test, Google engineers drive along the route one or more times to gather data about the environment. When it’s the autonomous vehicle’s turn to drive itself, it compares the data it is acquiring to the previously recorded data, an approach that is useful to differentiate pedestrians from stationary objects like poles and mailboxes.<<

This seems to require an amazing level of detail about roads to be known. And a lot of ongoing review required. All the time.

Whoops, that was thank you Crazy.

buffy said:

I don’t think I’ve actually got an answer yet. So far people have just said they can do it. How does a driverless car locate and read traffic lights, particularly at intersections where they are not directly across (ie, not 90/180 degree crosses) and where there are arrows?

video recognition.

i’d guess a lot of detail regards locations of traffic lights would be accessed by the equipment. It’d be doing near field traffic analysis as it negotiates traffic.

It wouldn’t be anything like farm boy here visiting adelaide and the sensation jeesus there’s a lot of lights and signs, and lanes branching to god knows where.

the thing about a robot car is it doesn’t worry in the human sense about it’s lack of familiarity, or worry about nasty thoughts from other drivers.

buffy said:

Thank you diddly, that is the sort of information I was looking for. From the second link:

>> Two things seem particularly interesting about Google’s approach. First, it relies on very detailed maps of the roads and terrain, something that Urmson said is essential to determine accurately where the car is. Using GPS-based techniques alone, he said, the location could be off by several meters.

The second thing is that, before sending the self-driving car on a road test, Google engineers drive along the route one or more times to gather data about the environment. When it’s the autonomous vehicle’s turn to drive itself, it compares the data it is acquiring to the previously recorded data, an approach that is useful to differentiate pedestrians from stationary objects like poles and mailboxes.<<

This seems to require an amazing level of detail about roads to be known. And a lot of ongoing review required. All the time.

I think people have been underestimating the complexity of self-driven traffic and the vast capacity for tragic cockups.

poikilotherm said:

buffy said:

I don’t think I’ve actually got an answer yet. So far people have just said they can do it. How does a driverless car locate and read traffic lights, particularly at intersections where they are not directly across (ie, not 90/180 degree crosses) and where there are arrows?

video recognition.

4 radars, two front, two at the back
Video
a Velodyne 64-beam laser, to generate detailed 3D environment
maps
GPS
intelligent cruise control,
ultrasonic sensors for wheel encoding
altimeters,
gyroscopes
and a tachometer

the second link I posted as a good video about it

Bubblecar said:

buffy said:

Thank you diddly, that is the sort of information I was looking for. From the second link:

>> Two things seem particularly interesting about Google’s approach. First, it relies on very detailed maps of the roads and terrain, something that Urmson said is essential to determine accurately where the car is. Using GPS-based techniques alone, he said, the location could be off by several meters.

The second thing is that, before sending the self-driving car on a road test, Google engineers drive along the route one or more times to gather data about the environment. When it’s the autonomous vehicle’s turn to drive itself, it compares the data it is acquiring to the previously recorded data, an approach that is useful to differentiate pedestrians from stationary objects like poles and mailboxes.<<

This seems to require an amazing level of detail about roads to be known. And a lot of ongoing review required. All the time.

I think people have been underestimating the complexity of self-driven traffic and the vast capacity for tragic cockups.

I think a human system is orders of magnitude more complex than than a self-driven system.

diddly-squat said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

The problem being that there is no computer system anywhere that is capable of interpreting the jumble of random lights that occur at many large intersections, let alone one that is small and cheap enough to build into a driverless car.

Safe driverless cars on general purpose urban roads are a long way off.

diddly-squat said:

I think a human system is orders of magnitude more complex than than a self-driven system.

Which is precisely why coping with general traffic is far more difficult than many people seem to realise.

The Rev Dodgson said:

diddly-squat said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

The problem being that there is no computer system anywhere that is capable of interpreting the jumble of random lights that occur at many large intersections, let alone one that is small and cheap enough to build into a driverless car.

Safe driverless cars on general purpose urban roads are a long way off.

The systems are getting much better, no doubt there will evolve an integrated system where signs etc are signed or chipped in a way that car systems can read them. Might be some sort of navigation alert “ you are now entering an autonomous equipped road network and your car may take full control” or similar.

The Rev Dodgson said:

diddly-squat said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

The problem being that there is no computer system anywhere that is capable of interpreting the jumble of random lights that occur at many large intersections, let alone one that is small and cheap enough to build into a driverless car.

Safe driverless cars on general purpose urban roads are a long way off.

um, there are several governments passing laws on driver less cars

which are here now

http://www.businessinsider.com.au/south-australia-has-just-legalised-driverless-cars-2016-4

The Rev Dodgson said:

diddly-squat said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

The problem being that there is no computer system anywhere that is capable of interpreting the jumble of random lights that occur at many large intersections, let alone one that is small and cheap enough to build into a driverless car.

Safe driverless cars on general purpose urban roads are a long way off.

Wouldn’t most of the jumble of lights not apply to the car in question though, you’d only be concerned with the ones related to giving the car the signal to stop, go or slow down and when it can or cannot turn

CrazyNeutrino said:

The Rev Dodgson said:

diddly-squat said:

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

The problem being that there is no computer system anywhere that is capable of interpreting the jumble of random lights that occur at many large intersections, let alone one that is small and cheap enough to build into a driverless car.

Safe driverless cars on general purpose urban roads are a long way off.

um, there are several governments passing laws on driver less cars

which are here now

http://www.businessinsider.com.au/south-australia-has-just-legalised-driverless-cars-2016-4

South Australia has just legalised driverless cars

South Australia has become the first Australian state to pass legislation around driverless cars.

The new laws will allow manufacturers to test autonomous vehicles on South Australian roads, only needing to apply for approval from the state transport minister to get them on the road.

“By being the first state in Australia to pass these laws we are sending a very clear message to this industry that South Australia is open for business,” SA transport minister Stephen Mulligan said.

CrazyNeutrino said:

The Rev Dodgson said:

diddly-squat said:

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

The problem being that there is no computer system anywhere that is capable of interpreting the jumble of random lights that occur at many large intersections, let alone one that is small and cheap enough to build into a driverless car.

Safe driverless cars on general purpose urban roads are a long way off.

um, there are several governments passing laws on driver less cars

which are here now

http://www.businessinsider.com.au/south-australia-has-just-legalised-driverless-cars-2016-4

That legislation makes driverless cars safe on general purpose urban roads does it?

The Rev Dodgson said:

CrazyNeutrino said:

The Rev Dodgson said:

The problem being that there is no computer system anywhere that is capable of interpreting the jumble of random lights that occur at many large intersections, let alone one that is small and cheap enough to build into a driverless car.

Safe driverless cars on general purpose urban roads are a long way off.

um, there are several governments passing laws on driver less cars

which are here now

http://www.businessinsider.com.au/south-australia-has-just-legalised-driverless-cars-2016-4

That legislation makes driverless cars safe on general purpose urban roads does it?

No, research and testing does, but you know that

legislation allows driverless cars to use the roads

there might be areas where driverless cars are allowed or not allowed

or other restrictions put in place

Cymek said:

Wouldn’t most of the jumble of lights not apply to the car in question though, you’d only be concerned with the ones related to giving the car the signal to stop, go or slow down and when it can or cannot turn

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

CrazyNeutrino said:

there might be areas where driverless cars are allowed or not allowed

or other restrictions put in place

So you agree then. As of now, no driverless car is safe for use on general purpose roads in urban areas.

The Rev Dodgson said:

Cymek said:

Wouldn’t most of the jumble of lights not apply to the car in question though, you’d only be concerned with the ones related to giving the car the signal to stop, go or slow down and when it can or cannot turn

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

Field of view and pattern recognition possibly

The Rev Dodgson said:

Cymek said:

Wouldn’t most of the jumble of lights not apply to the car in question though, you’d only be concerned with the ones related to giving the car the signal to stop, go or slow down and when it can or cannot turn

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

have a look at this youtube video

How Google’s Self-Driving Car Works
https://www.youtube.com/watch?v=YXylqtEQ0tk

How Google’s Self-Driving Car Works

The Rev Dodgson said:

Cymek said:

Wouldn’t most of the jumble of lights not apply to the car in question though, you’d only be concerned with the ones related to giving the car the signal to stop, go or slow down and when it can or cannot turn

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

So far I think we have got as far as:

There are all sorts of sensors and programs to read the sensors.

Google does lots and lots of very detailed mapping by people on the stretches of road they put the cars through.

The Rev Dodgson said:

CrazyNeutrino said:

there might be areas where driverless cars are allowed or not allowed

or other restrictions put in place

So you agree then. As of now, no driverless car is safe for use on general purpose roads in urban areas.

are we safe from drink drivers?

are we safe from people who use mobile phones while driving

what if the driverless cars turn out to be better at driving than humans

Cymek said:

The Rev Dodgson said:

Cymek said:

Wouldn’t most of the jumble of lights not apply to the car in question though, you’d only be concerned with the ones related to giving the car the signal to stop, go or slow down and when it can or cannot turn

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

Field of view and pattern recognition possibly

The problem is, computers don’t go around with a continually updating model of the world around them inside their memory chips.

Perhaps we can teach the driverless cars how to determine what signals/signs apply to it.

refit every set of traffic lights with radio transmitters.

CrazyNeutrino said:

The Rev Dodgson said:

Cymek said:

Wouldn’t most of the jumble of lights not apply to the car in question though, you’d only be concerned with the ones related to giving the car the signal to stop, go or slow down and when it can or cannot turn

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

have a look at this youtube video

How Google’s Self-Driving Car Works
https://www.youtube.com/watch?v=YXylqtEQ0tk

How Google’s Self-Driving Car Works

I know how self driving cars work.

That is the basis of my belief that they still have a long way to go before they will be safe on other than segregated roads with high standard signals.

Cymek said:

Perhaps we can teach the driverless cars how to determine what signals/signs apply to it.

Which might be what the Google engineers are doing with their intricate mappings. But the world changes. Minute by minute.

party_pants said:

refit every set of traffic lights with radio transmitters.

Trams and emergency service vehicles in Melbourne have transponders and can adjust the traffic lights for their needs.

Have a drive off with members of this forum and the winners brain is cloned and used in the driverless cars, aren’t member of this forum amongst the best drivers on the planet

CrazyNeutrino said:

The Rev Dodgson said:

CrazyNeutrino said:

there might be areas where driverless cars are allowed or not allowed

or other restrictions put in place

So you agree then. As of now, no driverless car is safe for use on general purpose roads in urban areas.

are we safe from drink drivers?

are we safe from people who use mobile phones while driving

what if the driverless cars turn out to be better at driving than humans

Drink driving and using mobile phones are both illegal.

The Rev Dodgson said:

Cymek said:

The Rev Dodgson said:

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

Field of view and pattern recognition possibly

The problem is, computers don’t go around with a continually updating model of the world around them inside their memory chips.

you missed the bit about the Velodyne 64-beam laser,

http://velodynelidar.com/hdl-64e.html

The HDL-64E LiDAR sensor is designed for obstacle detection and navigation of autonomous ground vehicles and marine vessels. Its durability, 360° field of view and very high data rate makes this sensor ideal for the most demanding perception applications as well as 3D mobile data collection and mapping applications. The HDL-64E’s innovative laser array enables navigation and mapping systems to observe more of their environment than any other LiDAR sensor.

buffy said:

Cymek said:

Perhaps we can teach the driverless cars how to determine what signals/signs apply to it.

Which might be what the Google engineers are doing with their intricate mappings. But the world changes. Minute by minute.

Roads less so perhaps

The Rev Dodgson said:

CrazyNeutrino said:

The Rev Dodgson said:

So you agree then. As of now, no driverless car is safe for use on general purpose roads in urban areas.

are we safe from drink drivers?

are we safe from people who use mobile phones while driving

what if the driverless cars turn out to be better at driving than humans

Drink driving and using mobile phones are both illegal.

What about falling asleep or not paying attention?

driverless cars do not do that

CrazyNeutrino said:

The Rev Dodgson said:

Cymek said:

Field of view and pattern recognition possibly

The problem is, computers don’t go around with a continually updating model of the world around them inside their memory chips.

you missed the bit about the Velodyne 64-beam laser,

http://velodynelidar.com/hdl-64e.html

The HDL-64E LiDAR sensor is designed for obstacle detection and navigation of autonomous ground vehicles and marine vessels. Its durability, 360° field of view and very high data rate makes this sensor ideal for the most demanding perception applications as well as 3D mobile data collection and mapping applications. The HDL-64E’s innovative laser array enables navigation and mapping systems to observe more of their environment than any other LiDAR sensor.

No I didn’t.

It’s not the sensors that are the problem. It’s what you do with all the data.

The Rev Dodgson said:

CrazyNeutrino said:

The Rev Dodgson said:

The problem is, computers don’t go around with a continually updating model of the world around them inside their memory chips.

you missed the bit about the Velodyne 64-beam laser,

http://velodynelidar.com/hdl-64e.html

The HDL-64E LiDAR sensor is designed for obstacle detection and navigation of autonomous ground vehicles and marine vessels. Its durability, 360° field of view and very high data rate makes this sensor ideal for the most demanding perception applications as well as 3D mobile data collection and mapping applications. The HDL-64E’s innovative laser array enables navigation and mapping systems to observe more of their environment than any other LiDAR sensor.

No I didn’t.

It’s not the sensors that are the problem. It’s what you do with all the data.

Did you look at the video?

https://www.youtube.com/watch?v=YXylqtEQ0tk

The Rev Dodgson said:

CrazyNeutrino said:

The Rev Dodgson said:

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

have a look at this youtube video

How Google’s Self-Driving Car Works
https://www.youtube.com/watch?v=YXylqtEQ0tk

How Google’s Self-Driving Car Works

I know how self driving cars work.

That is the basis of my belief that they still have a long way to go before they will be safe on other than segregated roads with high standard signals.

Why do they still have a long way to go?

CrazyNeutrino said:

The Rev Dodgson said:

CrazyNeutrino said:

are we safe from drink drivers?

are we safe from people who use mobile phones while driving

what if the driverless cars turn out to be better at driving than humans

Drink driving and using mobile phones are both illegal.

What about falling asleep or not paying attention?

driverless cars do not do that

Falling asleep and not paying attention are also illegal, but that’s not the point.

Computer systems do things that have consequences similar to that of falling asleep or not paying attention all the time.

Autonomous Car

In Europe

In Europe, cities in Belgium, France, Italy and the UK are planning to operate transport systems for driverless cars, and Germany, the Netherlands, and Spain have allowed testing robotic cars in traffic. In 2015, the UK Government launched public trials of the LUTZ Pathfinder driverless pod in Milton Keynes. Since Summer 2015 the French government allowed PSA Peugeot-Citroen to make trials in real conditions in the Paris area. The experiments will be extended to other French cities like Bordeaux and Strasbourg by 2016. The alliance between the French companies THALES and Valeo (provider of the first self-parking car system that equips Audi and Mercedes premi) is also testing its own driverless car system.

and in the US

To clarify the legal status of and otherwise regulate such vehicles, several states have enacted or are considering specific laws. In 2016, 7 states (Nevada, California, Florida, Michigan, Hawaii, Washington, and Tennessee), along with the District of Columbia, have enacted laws for autonomous vehicles.

The Rev Dodgson said:

CrazyNeutrino said:

The Rev Dodgson said:

Drink driving and using mobile phones are both illegal.

What about falling asleep or not paying attention?

driverless cars do not do that

Falling asleep and not paying attention are also illegal, but that’s not the point.

Computer systems do things that have consequences similar to that of falling asleep or not paying attention all the time.

You’ve just got to turn it off. And turn it on again.

:)

CrazyNeutrino said:

Why do they still have a long way to go?

Because:

1) Computer systems dealing with complex situations are highly unreliable.

2) No computer system, even the largest and most advanced super-computers, generate, continually update, and understand an internal model of the world, in the same way that humans do. It is that modelling that allows us to deal with unexpected situations quickly and relatively safely.

>are we safe from people who use mobile phones while driving

this inattention to primary task proposition, needs be understood in the context of driver behaviours that tend unsafe proximity, meaning some aspects of high density traffic incline regular breaches of safe distances.

>what if the driverless cars turn out to be better at driving than humans

what you’ve done is shift your question into an undefined “better”.

The Rev Dodgson said:

CrazyNeutrino said:

The Rev Dodgson said:

That was the question. How does the car differentiate the traffic lights that apply to it, from all the other traffic lights, and all the other non-traffic lights on a general purpose road with standard traffic lights?

have a look at this youtube video

How Google’s Self-Driving Car Works
https://www.youtube.com/watch?v=YXylqtEQ0tk

How Google’s Self-Driving Car Works

I know how self driving cars work.

That is the basis of my belief that they still have a long way to go before they will be safe on other than segregated roads with high standard signals.

A quick googling reveals something amusing, many articles about the ‘death of traffic lights’ with driverless cars…then there’s one article about how much self driving cars struggle with traffic lights.

Perhaps we pretend they are driverless but hide a little person between the dashboard and the engine bay and they do the driving

CrazyNeutrino said:

Did you look at the video?

https://www.youtube.com/watch?v=YXylqtEQ0tk

I’m not sitting through that.

The Rev Dodgson said:

CrazyNeutrino said:

Did you look at the video?

https://www.youtube.com/watch?v=YXylqtEQ0tk

I’m not sitting through that.

You could always stand instead

The Rev Dodgson said:

CrazyNeutrino said:

Did you look at the video?

https://www.youtube.com/watch?v=YXylqtEQ0tk

I’m not sitting through that.

its good information

the first speaker is hard to understand but the second speaker is ok

skip forward to 2:20 and start from there

The fidelity of vision must be pretty good. I was reading about a system that got confused by a cyclist doing that forward and backward and thing they do to maintain balance. At the intersection the computer couldn’t figure out if it was going to enter or not. Google engineers were with it and logged it an another thing the sytems have to be aware of.

CrazyNeutrino said:

The Rev Dodgson said:

CrazyNeutrino said:

Did you look at the video?

https://www.youtube.com/watch?v=YXylqtEQ0tk

I’m not sitting through that.

its good information

the first speaker is hard to understand but the second speaker is ok

skip forward to 2:20 and start from there

Sorry, but it doesn’t address the issue.

I haven’t watched it (other than the first 2 minutes), but I’m 99.9% confident that it doesn’t address the issue.

Anyway, I’ve seen these things before.

The first one was in 1972 I think, possibly earlier.

The Rev Dodgson said:

CrazyNeutrino said:

The Rev Dodgson said:

I’m not sitting through that.

its good information

the first speaker is hard to understand but the second speaker is ok

skip forward to 2:20 and start from there

Sorry, but it doesn’t address the issue.

I haven’t watched it (other than the first 2 minutes), but I’m 99.9% confident that it doesn’t address the issue.

Anyway, I’ve seen these things before.

The first one was in 1972 I think, possibly earlier.

When driverless cars are common

look back at this thread

CrazyNeutrino said:

When driverless cars are common

look back at this thread

How about responding to what I say, rather than repeating the same stuff that doesn’t?

Anyway, going to do something useful now.

The Rev Dodgson said:

CrazyNeutrino said:

When driverless cars are common

look back at this thread

How about responding to what I say, rather than repeating the same stuff that doesn’t?

Anyway, going to do something useful now.

Did you watch the video from 2:20?

besides all the technology each state in Australia has some kind of Road traffic authority, which has driving experts, the driver less cars would have to please the road traffic authority to pass inspection and allow driver less cars to operate

Driverless cars would go through a phase of RTA testing etc

CrazyNeutrino said:

besides all the technology each state in Australia has some kind of Road traffic authority, which has driving experts, the driver less cars would have to please the road traffic authority to pass inspection and allow driver less cars to operate

Driverless cars would go through a phase of RTA testing etc

Plus the driverless car could be identifiable by the general public and they show patience, care and understanding like they do for all other human drivers on the road.

CrazyNeutrino said:

besides all the technology each state in Australia has some kind of Road traffic authority, which has driving experts, the driver less cars would have to please the road traffic authority to pass inspection and allow driver less cars to operate

Driverless cars would go through a phase of RTA testing etc

All the road rule boxes would have to be ticked, I would imagine that driverless cars and trucks are in a “work in progress stage” and will be for some time.

Cymek said:

CrazyNeutrino said:

besides all the technology each state in Australia has some kind of Road traffic authority, which has driving experts, the driver less cars would have to please the road traffic authority to pass inspection and allow driver less cars to operate

Driverless cars would go through a phase of RTA testing etc

Plus the driverless car could be identifiable by the general public and they show patience, care and understanding like they do for all other human drivers on the road.

they may have special number plates

or have special colorings like taxis, taxis are easily identifiable t

they may transmit a wifi signal indicating driverless

Cymek said:

CrazyNeutrino said:

besides all the technology each state in Australia has some kind of Road traffic authority, which has driving experts, the driver less cars would have to please the road traffic authority to pass inspection and allow driver less cars to operate

Driverless cars would go through a phase of RTA testing etc

Plus the driverless car could be identifiable by the general public and they show patience, care and understanding like they do for all other human drivers on the road.

I think you needed a tic on that one Cymek. It got lost in translation.

The Rev Dodgson said:

diddly-squat said:

buffy said:

While we were driving around in Melbourne over the weekend, a question occurred to me. How would driverless cars cope with those large intersections controlled by a conglomeration of traffic lights and red and green arrows and stuff? We considered that they might take their cue from the control boxes, but I doubt the infrastructure is really set up for that.

they would simply follow the rules of the road; that is they would wait for a green light and proceed accordingly – or in the absence of lights, they would follow standard give way rules and then proceed accordingly.

There would be no need for any external controlling mechanism.

The problem being that there is no computer system anywhere that is capable of interpreting the jumble of random lights that occur at many large intersections, let alone one that is small and cheap enough to build into a driverless car.

Safe driverless cars on general purpose urban roads are a long way off.

we have the technology now… it’s been proven.. driverless cars have driven tens of thousands of hours unaided

Is Tony Abbott is using Revs handle?

give it back now

i’d think driverless cars are technologically quite doable presently, however not with typical speeds and traffic densities that exist, or to put it another way having many driverless vehicles on the roads will change the typical traffic densities because a lot of driving is about proximity between vehicles tending to regulate speed(and more). The behaviour (driving) is about fitting in, some of the surrounds is static, some is moving.

the driverless vehicles will become part of what humans drivers fit into. Further, down the track the driverless vehicles will become part of the policing presence on the road, and they’ll somewhat replace existing police on the road. They need not be government either, they could be private. So there’re people with money out there that’d be looking to invest in that.

for Rev

33 Corporations Working On Autonomous Vehicles
https://www.cbinsights.com/blog/autonomous-driverless-vehicles-corporations-list/

33 Corporations Working On Autonomous Vehicles

here is a google self driving car. it looks distinctive

I suspect making them obvious is not necessarily a good idea. Roadcraft is not a widely practised thing.

>here is a google self driving car. it looks distinctive

oneday there will be vehicles everywhere with radomes atop, the mil’ GPS guided machines with eyes helping humans efficiently get A-B

The Ad for the Mercedes F 015 D
https://www.mercedes-benz.com/en/mercedes-benz/innovation/research-vehicle-f-015-luxury-in-motion/

Mercedes F 015 Drives Itself To CES Las Vegas
https://www.youtube.com/watch?v=VfbJYWrXS9I

Mercedes Auto Truck demo
https://www.youtube.com/watch?v=eQ5tUI1xW8E

Mercedes F 015 World Premiere In Detail Walk Around Mercedes Self Driving Car CARJAM TV 2015

Audi Self Driving Car A9 Highway Demo A7 Audi Piloted Driving How It Works CARJAM TV HD 2016

Range Rover Self Driving Car Off Road World First Jaguar Auto Pilot Autonomous Car 2016 CARJAM TV

VW Self Driving Car VW Self Parking Car 2017 VW eGolf Electric V Charge VW Commercial CARJAM TV HD

I think it is better to think of the car standing still until all the requirements for moving forward are met rather than thinking of the car as being okay to proceed until it meets a red light.

2 different BMW self driving cars

the BMW 7 Series
The innovative features of the all new BMW 7 Series Self Driving Car

the BMW 3 Series
https://www.youtube.com/watch?v=WsnKzK6dX8Q

Volvo Pilot Assist – Low-Speed Autonomous Driving Review

CrazyNeutrino said:

2 different BMW self driving cars

the BMW 7 Series
The innovative features of the all new BMW 7 Series Self Driving Car

the BMW 3 Series
https://www.youtube.com/watch?v=WsnKzK6dX8Q

former top gear jeremy tests the 3 series

Rolls-Royce Vision World Premiere Review Rolls Royce Vision Self Driving Car CARJAM TV HD 2016

TESLA AUTOPILOT ROCKS !!

Could you refine your search a bit more please Crazy and find out how the systems differentiate the appropriate traffic light from the visual noise.

My question was not are there lots of companies spruiking their research. I asked a very specific question.

buffy said:

Could you refine your search a bit more please Crazy and find out how the systems differentiate the appropriate traffic light from the visual noise.

My question was not are there lots of companies spruiking their research. I asked a very specific question.

It is possible, right now, to walk into a camera shop and buy a very inexpensive camera that has inbuilt facial recognition capability – current generation smart phones can do it also. It’s not a particularly big leap, from there, to imagine a camera that has the ability to ‘recognise’ a stop light and act accordingly.

Facebook has the functionality to scan pictures for people you know and automatically tag them.

buffy said:

Could you refine your search a bit more please Crazy and find out how the systems differentiate the appropriate traffic light from the visual noise.

My question was not are there lots of companies spruiking their research. I asked a very specific question.

I guess its a bit like face finding tech you see on point and shoot cameras but more complicated

here is an article on Googles Traffic Light detection system

Traffic Light Mapping and Detection

>Could you refine your search a bit more please Crazy and find out how the systems differentiate the appropriate traffic light from the visual noise.

buffy’s worried that if she’s standing at a walk crossing with her favourite green dress on an autonomous vehicle is going to pull out and cause chaos.

diddly-squat said:

It is possible, right now, to walk into a camera shop and buy a very inexpensive camera that has inbuilt facial recognition capability – current generation smart phones can do it also. It’s not a particularly big leap, from there, to imagine a camera that has the ability to ‘recognise’ a stop light and act accordingly.

Facebook has the functionality to scan pictures for people you know and automatically tag them.

The trouble is, face recognition under anything but controlled conditions is absolutely hopeless. Even under controlled conditions it is pretty slow.

Google also has functionality to scan for similar images. If there is another copy of the same image somewhere on the Internet it works just great. On the other hand if there is an image of the same thing taken from a slightly different angle, or with different lighting, it is hopeless. More often than not it fails to find any images of what you are looking for, and comes up with a whole load of suggestions of things that are quite obviously something totally different. But it doesn’t know that; it’s comparison algorithm comes up with a score that suggests they might be the same thing.

For those unfamiliar with Lidar

Lidar is a detection system which works on the principle of radar, but uses light from a laser.

They use this to get a 3D map of the cars external environment in real time

The Rev Dodgson said:

diddly-squat said:

It is possible, right now, to walk into a camera shop and buy a very inexpensive camera that has inbuilt facial recognition capability – current generation smart phones can do it also. It’s not a particularly big leap, from there, to imagine a camera that has the ability to ‘recognise’ a stop light and act accordingly.

Facebook has the functionality to scan pictures for people you know and automatically tag them.

The trouble is, face recognition under anything but controlled conditions is absolutely hopeless. Even under controlled conditions it is pretty slow.

Google also has functionality to scan for similar images. If there is another copy of the same image somewhere on the Internet it works just great. On the other hand if there is an image of the same thing taken from a slightly different angle, or with different lighting, it is hopeless. More often than not it fails to find any images of what you are looking for, and comes up with a whole load of suggestions of things that are quite obviously something totally different. But it doesn’t know that; it’s comparison algorithm comes up with a score that suggests they might be the same thing.

Have a look at Googles Traffic Light Mapping and Detection System

CrazyNeutrino said:

buffy said:

Could you refine your search a bit more please Crazy and find out how the systems differentiate the appropriate traffic light from the visual noise.

My question was not are there lots of companies spruiking their research. I asked a very specific question.

I guess its a bit like face finding tech you see on point and shoot cameras but more complicated

here is an article on Googles Traffic Light detection system

Traffic Light Mapping and Detection

Finally something with some answers to Buffy’s question (albeit written by Google).

In summary, at the moment to do anything like a reasonable job they need to map all the lights in detail. They are still working on it.

CrazyNeutrino said:

The Rev Dodgson said:

diddly-squat said:

It is possible, right now, to walk into a camera shop and buy a very inexpensive camera that has inbuilt facial recognition capability – current generation smart phones can do it also. It’s not a particularly big leap, from there, to imagine a camera that has the ability to ‘recognise’ a stop light and act accordingly.

Facebook has the functionality to scan pictures for people you know and automatically tag them.

The trouble is, face recognition under anything but controlled conditions is absolutely hopeless. Even under controlled conditions it is pretty slow.

Google also has functionality to scan for similar images. If there is another copy of the same image somewhere on the Internet it works just great. On the other hand if there is an image of the same thing taken from a slightly different angle, or with different lighting, it is hopeless. More often than not it fails to find any images of what you are looking for, and comes up with a whole load of suggestions of things that are quite obviously something totally different. But it doesn’t know that; it’s comparison algorithm comes up with a score that suggests they might be the same thing.

Have a look at Googles Traffic Light Mapping and Detection System

Have you read that?

The Rev Dodgson said:

CrazyNeutrino said:

The Rev Dodgson said:

The trouble is, face recognition under anything but controlled conditions is absolutely hopeless. Even under controlled conditions it is pretty slow.

Google also has functionality to scan for similar images. If there is another copy of the same image somewhere on the Internet it works just great. On the other hand if there is an image of the same thing taken from a slightly different angle, or with different lighting, it is hopeless. More often than not it fails to find any images of what you are looking for, and comes up with a whole load of suggestions of things that are quite obviously something totally different. But it doesn’t know that; it’s comparison algorithm comes up with a score that suggests they might be the same thing.

Have a look at Googles Traffic Light Mapping and Detection System

Have you read that?

Yes, Im still reading it

its 6 pages long

How does a driver-less car read the traffic lights when stuck behind a big truck?

CrazyNeutrino said:

The Rev Dodgson said:

CrazyNeutrino said:

Have a look at Googles Traffic Light Mapping and Detection System

Have you read that?

Yes, Im still reading it

its 6 pages long

http://static.googleusercontent.com/media/research.google.com/en//pubs/archive/37259.pdf

We use a fixed lens with a 30 degree field of view, which we selected with the goal of sufficient resolution to be able to detect traffic lights out to 150 m, a reasonable braking distance when traveling at 55mph Since our detector depends primarily on color because no structure is visible at night, we also fix the gain and shutter speeds to avoid saturation of the traffic lights, particularly bright LED-based green lights.

PermeateFree said:

How does a driver-less car read the traffic lights when stuck behind a big truck?

Google are using predicative mapping as well as a range of sensors that operate in real time

Another idea is to use wifi on traffic lights, a wifi signal can indicate which state the traffic light is in ,and the duration of each state, stop, caution and go

smart cars could use those wifi signals to alter their speed and drive though more green lights

another idea is go get rid of traffic lights altogether

but in the mean time we use existing technologies

note that traffic lights usually have a specific shape

also their have specific colors

LED traffic lights can be made uniform and consistent to have specific colors for red yellow and green

these can be specific to one frequencies or within a small range of frequencies

the color itself is another way to detect the traffic lights in addition to its mapped location

So in short, it’s a work in progress and they still have a load of problems to sort out.

The Rev Dodgson said:

So in short, it’s a work in progress and they still have a load of problems to sort out.

Yes

The Rev Dodgson said:

So in short, it’s a work in progress and they still have a load of problems to sort out.

but bear in mind that various governments have passed driver less laws and there are various forms of driver-less cars already out there

the Tesla in the above link being a good example

The Rev Dodgson said:

diddly-squat said:

It is possible, right now, to walk into a camera shop and buy a very inexpensive camera that has inbuilt facial recognition capability – current generation smart phones can do it also. It’s not a particularly big leap, from there, to imagine a camera that has the ability to ‘recognise’ a stop light and act accordingly.

Facebook has the functionality to scan pictures for people you know and automatically tag them.

The trouble is, face recognition under anything but controlled conditions is absolutely hopeless. Even under controlled conditions it is pretty slow.

Google also has functionality to scan for similar images. If there is another copy of the same image somewhere on the Internet it works just great. On the other hand if there is an image of the same thing taken from a slightly different angle, or with different lighting, it is hopeless. More often than not it fails to find any images of what you are looking for, and comes up with a whole load of suggestions of things that are quite obviously something totally different. But it doesn’t know that; it’s comparison algorithm comes up with a score that suggests they might be the same thing.

maybe I’m missing something but when I use my camera it certainly doesn’t have any problems finding where people’s faces are…

PermeateFree said:

How does a driver-less car read the traffic lights when stuck behind a big truck?

why would it have to? it’s stuck behind a truck…

diddly-squat said:

The Rev Dodgson said:

diddly-squat said:

It is possible, right now, to walk into a camera shop and buy a very inexpensive camera that has inbuilt facial recognition capability – current generation smart phones can do it also. It’s not a particularly big leap, from there, to imagine a camera that has the ability to ‘recognise’ a stop light and act accordingly.

Facebook has the functionality to scan pictures for people you know and automatically tag them.

The trouble is, face recognition under anything but controlled conditions is absolutely hopeless. Even under controlled conditions it is pretty slow.

Google also has functionality to scan for similar images. If there is another copy of the same image somewhere on the Internet it works just great. On the other hand if there is an image of the same thing taken from a slightly different angle, or with different lighting, it is hopeless. More often than not it fails to find any images of what you are looking for, and comes up with a whole load of suggestions of things that are quite obviously something totally different. But it doesn’t know that; it’s comparison algorithm comes up with a score that suggests they might be the same thing.

maybe I’m missing something but when I use my camera it certainly doesn’t have any problems finding where people’s faces are…

it has face finding technology

traffic lights have three distinct colrs and each has a distant shape, there are round

so three lights in a vertical row

using shape and color detection along with predicative mapping of where the traffic lights should be

different manufacturers are using slightly different technique’s

so laws will have to presumably cover all the different techniques

some maybe better than others

FB manages to identify my family members under conditions that are not controlled in any way so I think Rev’s assessment is a bit pessimistic.

dv said:

FB manages to identify my family members under conditions that are not controlled in any way so I think Rev’s assessment is a bit pessimistic.

I think his assessment of the capabilities of driverless cars is broadly pessimistic

diddly-squat said:

The Rev Dodgson said:

diddly-squat said:

It is possible, right now, to walk into a camera shop and buy a very inexpensive camera that has inbuilt facial recognition capability – current generation smart phones can do it also. It’s not a particularly big leap, from there, to imagine a camera that has the ability to ‘recognise’ a stop light and act accordingly.

Facebook has the functionality to scan pictures for people you know and automatically tag them.

The trouble is, face recognition under anything but controlled conditions is absolutely hopeless. Even under controlled conditions it is pretty slow.

Google also has functionality to scan for similar images. If there is another copy of the same image somewhere on the Internet it works just great. On the other hand if there is an image of the same thing taken from a slightly different angle, or with different lighting, it is hopeless. More often than not it fails to find any images of what you are looking for, and comes up with a whole load of suggestions of things that are quite obviously something totally different. But it doesn’t know that; it’s comparison algorithm comes up with a score that suggests they might be the same thing.

maybe I’m missing something but when I use my camera it certainly doesn’t have any problems finding where people’s faces are…

What about these people ?

Have a look at this Video of a Tesla with Autodrive

TESLA AUTOPILOT Model S Badmitton

diddly-squat said:

PermeateFree said:

How does a driver-less car read the traffic lights when stuck behind a big truck?

why would it have to? it’s stuck behind a truck…

More to the point, how do YOU read the traffic lights when stuck behind a big truck?
Driverless cars have more eyes than you.

Longer Version
Tesla Autopilot in P85D Supercar, Scary as Hell!
https://www.youtube.com/watch?v=J6905HDRs04

This whole thread reminds me of something Wernher von Braun once said:
Wernher von Braun said:

Man is the best computer we can put aboard a spacecraft … and the only one that can be mass-produced with unskilled labour.

one thing I do know will be a welcome development of autonomous vehicles is that it can change a flat tyre unassisted, as this development will turn up on regular cars.

PermeateFree said:

How does a driver-less car read the traffic lights when stuck behind a big truck?

in a sense brake lights are stop lights, so vehicles ahead (of traffic flow and behaviour) serve as something similar to stop lights

it’s worth a thought in thinking about how different are pole mounted stoplights.

transition said:

one thing I do know will be a welcome development of autonomous vehicles is that it can change a flat tyre unassisted, as this development will turn up on regular cars.

Doesn’t seem to be much of an issue these days due to improved tyre technology. I have had no cause to change a tyre on the road since the early 1990s.

Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike. Kew Junction in Melbourne, coming in from the East, you are not actually facing a road. You are facing the corner of a pub. There are four roads you could turn into (although one you aren’t allowed to). You have to discriminate which green light/arrow is the one applicable to the road you want to take. And they don’t all go at once. Similarly coming in from Footscray/Flemington Rd towards the Melbourne zoo. A doozy of an intersection. Both of these are complicated by tramtracks as well. I’m quite happy to accept that lights can be recognized. But the next step is to know which ones are yours. This is a far, far more difficult thing.

buffy said:

Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike.

(Shrugs) How do YOU do it?

dv said:

buffy said:

Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike.

(Shrugs) How do YOU do it?

Human brain. Far better discrimination and pattern recognition than any computer to date.

We are the pattern recognition kings.

dv said:

transition said:

one thing I do know will be a welcome development of autonomous vehicles is that it can change a flat tyre unassisted, as this development will turn up on regular cars.

Doesn’t seem to be much of an issue these days due to improved tyre technology. I have had no cause to change a tyre on the road since the early 1990s.

i’ve been fairly lucky too, been driving without a spare for a year, but I do know a mature lady that owns a particular brand of vehicle, that had the engineers set out to make a flat as difficult as possible to change it wouldn’t have been worse than it is. And she’s had a few flats.

buffy said:

dv said:

buffy said:

Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike.

(Shrugs) How do YOU do it?

Human brain. Far better discrimination and pattern recognition than any computer to date.

We are the pattern recognition kings.

When a computer can see the face of Jesus in a Dorito us humans will be obsolete

buffy said:

dv said:

buffy said:

Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike.

(Shrugs) How do YOU do it?

Human brain. Far better discrimination and pattern recognition than any computer to date.

We are the pattern recognition kings.

Ah well, I don’t agree with you that it would be difficult to program a computer to work this out, compared to the task of getting it to maneuvre safely in traffic. First, identify it as a fiveways, so anticipate five sets of lights facing in five directions, work out which one is suitable to the purpose, confirm by checking road markings.

>>Ah well, I don’t agree with you that it would be difficult to program a computer to work this out, compared to the task of getting it to maneuvre safely in traffic. <<

Manoeuvring in traffic is simple proximity sensing. Choosing the right traffic light is another step. You not only have to find it, you have to read it.

dv said:

buffy said:

dv said:

(Shrugs) How do YOU do it?

Human brain. Far better discrimination and pattern recognition than any computer to date.

We are the pattern recognition kings.

Ah well, I don’t agree with you that it would be difficult to program a computer to work this out, compared to the task of getting it to maneuvre safely in traffic. First, identify it as a fiveways, so anticipate five sets of lights facing in five directions, work out which one is suitable to the purpose, confirm by checking road markings.

I’ve used that particular intersection frequently, and so I’m used to it – at least from the directions I usually come from (north or south), but when you first see it you can take some time to figure out what signals apply to which streets. I’ve seen drivers obey the wrong signal, obviously mistaking it as applying to them.

As an aside, if you’re a cyclist travelling north in a head wind, you’ve got additional challenges because the applicable lights don’t last long enough for you to get across the street. But that doesn’t apply to driverless cars.

buffy said:

Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike. Kew Junction in Melbourne, coming in from the East, you are not actually facing a road. You are facing the corner of a pub. There are four roads you could turn into (although one you aren’t allowed to). You have to discriminate which green light/arrow is the one applicable to the road you want to take. And they don’t all go at once. Similarly coming in from Footscray/Flemington Rd towards the Melbourne zoo. A doozy of an intersection. Both of these are complicated by tramtracks as well. I’m quite happy to accept that lights can be recognized. But the next step is to know which ones are yours. This is a far, far more difficult thing.

I really think these things will need to be mapped in detail, and kept up to date (including am/pm changes in lane directions).

Just within 5 km of where I live there are two sets of lights where lights turn green for another traffic flow, and are visible to traffic stopped at red lights. So people who don’t know the lights start to drive off, then notice their lights are still red, and stop again. I’m not convinced that computers would deal with that sort of situation reliably, unless it was all programmed in and kept up to date somehow.

>Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike.

familiarization I suppose.

dv said:

buffy said:

Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike.

(Shrugs) How do YOU do it?

By referring to the model of the outside world inside your head, and making intelligent inferences based on your knowledge of how things all fit together.

Computers don’t have that sort of model.

transition said:

>Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike.

familiarization I suppose.

I think it’s called programming..sorry, coding, these days.

buffy said:

Alright, now we seem to have established that they can recognize a traffic light, but I still don’t understand how they can discriminate which one is applicable to them at fiveways intersections and suchlike. Kew Junction in Melbourne, coming in from the East, you are not actually facing a road. You are facing the corner of a pub. There are four roads you could turn into (although one you aren’t allowed to). You have to discriminate which green light/arrow is the one applicable to the road you want to take. And they don’t all go at once. Similarly coming in from Footscray/Flemington Rd towards the Melbourne zoo. A doozy of an intersection. Both of these are complicated by tramtracks as well. I’m quite happy to accept that lights can be recognized. But the next step is to know which ones are yours. This is a far, far more difficult thing.

one system they use is prior mapping of where a traffic light intersection should be

in other words they use data from a previous visit through the intersection

From
Traffic Light Mapping and Detection
http://static.googleusercontent.com/media/research.google.com/en//pubs/archive/37259.pdf

VII. CONCLUSIONS

Cars must deal with traffic lights. The two main tasks are
detecting the traffic lights and understanding their control
semantics. Our approach to solving these two tasks has been
to automatically construct maps of the traffic light positions
and orientations, and then to manually add control semantics
to each light.

We then use this map to allow an onboard
perception system to anticipate when it should see and react
to a traffic light, to improve the performance of the traffic
light detector by predicting the precise location of the traffic
light in the camera image, and to then determine whether
a particular route through an intersection is allowed. Our
system has been deployed on multiple cars, and has provided
reliable and timely information about the state of the traffic
lights during thousands of drives through intersections.

We are now experimenting with a secondary camera with
a wider field of view but a shorter detection range, that
allows the car to detect nearby lights when it is very close to
the intersection. We are also experimenting with the robust
detection of flashing lights: we can annotate the map with
lights that always flash or lights that flash at certain times
of day, but during construction or an emergency lights may
unpredictably be switched to flash. Finally, there are cases
where the lights, frequently arrows, are so dim, it appears
that the only way to detect the state is to watch for relative
changes in intensity of the light elements.

Thanks to Anna Osepayshvili for her analysis of the
human QC time required to build the traffic light maps

Traffic light identification could be enhanced by senors looking for specific frequency colors for Red, Yellow and Green

A standard could be set for all traffic lights around Australia

Manufacturers of traffic lights then adhere to the standard

In addition wifi signals could be sent from smart traffic lights that transmit their exact location and status and length of each state for stop, caution as well as the type of traffic light and go giving auto cars more information to guide them through an intersection or pedestrian lights, train lights, tram lights etc

>I think it’s called programming..sorry, coding, these days.

i’d say familiarization, of learning, is something a lot of the animal world does (not just humans), and that of manmade non organic machines same applies (they can do it, or be made to do it). Not sure it’s unique to replicators (a requirement), it maybe.

i’d guess something that builds representations of the environment, that has goals or objectives, that then implements ways (even or including try and test) of negotiating obstacles, involved or requires computation.

so I dunno, I doubt the concept of familiarization (the work of) is so different to computation. It just doesn’t sound trendy.

trendy’s the latest thing.

Prior learning of traffic light locations is involved with prior mapping

I guess eventually the system will develop into learning on the fly as well

could road safety be enhanced by cars having a little traffic light system where the ordinary rear lights are
red for stop
yellow for braking
green for go
as well as a digital speed indicator on the rear ?

CrazyNeutrino said:

could road safety be enhanced by cars having a little traffic light system where the ordinary rear lights are
red for stop
yellow for braking
green for go
as well as a digital speed indicator on the rear ?

or leave red for braking
but have a dual red for brake and stopped
green for go

?

or just leave it all untouched

CrazyNeutrino said:

From
Traffic Light Mapping and Detection
http://static.googleusercontent.com/media/research.google.com/en//pubs/archive/37259.pdf

VII. CONCLUSIONS

Cars must deal with traffic lights. The two main tasks are
detecting the traffic lights and understanding their control
semantics. Our approach to solving these two tasks has been
to automatically construct maps of the traffic light positions
and orientations, and then to manually add control semantics
to each light.

We then use this map to allow an onboard
perception system to anticipate when it should see and react
to a traffic light, to improve the performance of the traffic
light detector by predicting the precise location of the traffic
light in the camera image, and to then determine whether
a particular route through an intersection is allowed. Our
system has been deployed on multiple cars, and has provided
reliable and timely information about the state of the traffic
lights during thousands of drives through intersections.

We are now experimenting with a secondary camera with
a wider field of view but a shorter detection range, that
allows the car to detect nearby lights when it is very close to
the intersection. We are also experimenting with the robust
detection of flashing lights: we can annotate the map with
lights that always flash or lights that flash at certain times
of day, but during construction or an emergency lights may
unpredictably be switched to flash. Finally, there are cases
where the lights, frequently arrows, are so dim, it appears
that the only way to detect the state is to watch for relative
changes in intensity of the light elements.

Thanks to Anna Osepayshvili for her analysis of the
human QC time required to build the traffic light maps

This is more like what I was after. Mapping, mapping and more mapping. And then checking all the time – all the time – for changes to the road situation. It’s not an easy thing at all.

A traffic light wifi signal could indicate braking information to a whole row of cars

That research comes down to knowing everything about every intersection the car is going to go through.

one thing that’s easy to forget with the glorification of vehicular transport, is that a lot of people are not driving any more than they need, because driving, being out on the road is dangerous.

the future machines will be even less inclined to tell you this, in fact they’ll be a fearless role model.

buffy said:

CrazyNeutrino said:

From
Traffic Light Mapping and Detection
http://static.googleusercontent.com/media/research.google.com/en//pubs/archive/37259.pdf

VII. CONCLUSIONS

Cars must deal with traffic lights. The two main tasks are
detecting the traffic lights and understanding their control
semantics. Our approach to solving these two tasks has been
to automatically construct maps of the traffic light positions
and orientations, and then to manually add control semantics
to each light.

We then use this map to allow an onboard
perception system to anticipate when it should see and react
to a traffic light, to improve the performance of the traffic
light detector by predicting the precise location of the traffic
light in the camera image, and to then determine whether
a particular route through an intersection is allowed. Our
system has been deployed on multiple cars, and has provided
reliable and timely information about the state of the traffic
lights during thousands of drives through intersections.

We are now experimenting with a secondary camera with
a wider field of view but a shorter detection range, that
allows the car to detect nearby lights when it is very close to
the intersection. We are also experimenting with the robust
detection of flashing lights: we can annotate the map with
lights that always flash or lights that flash at certain times
of day, but during construction or an emergency lights may
unpredictably be switched to flash. Finally, there are cases
where the lights, frequently arrows, are so dim, it appears
that the only way to detect the state is to watch for relative
changes in intensity of the light elements.

Thanks to Anna Osepayshvili for her analysis of the
human QC time required to build the traffic light maps

This is more like what I was after. Mapping, mapping and more mapping. And then checking all the time – all the time – for changes to the road situation. It’s not an easy thing at all.

yes its not easy, it involves past information, lots of present real time information as well as predicative information

buffy said:

That research comes down to knowing everything about every intersection the car is going to go through.

more information the better

CrazyNeutrino said:

buffy said:

That research comes down to knowing everything about every intersection the car is going to go through.

more information the better

I don’t reckon real roads are that controlled.

buffy said:

CrazyNeutrino said:

buffy said:

That research comes down to knowing everything about every intersection the car is going to go through.

more information the better

I don’t reckon real roads are that controlled.

Busy intersections can even be a hazard to human drivers

lots of stories on crashes at intersections from hell

buffy said:

That research comes down to knowing everything about every intersection the car is going to go through.

Not impossible, just difficult…but if it was easy…everyone would be making them.

some intersections in Melb CBD I avoid altogether

if I know an intersection is a bad one I will avoid it

CrazyNeutrino said:

some intersections in Melb CBD I avoid altogether

if I know an intersection is a bad one I will avoid it

what constitutes a ‘bad’ intersection?

stumpy_seahorse said:

CrazyNeutrino said:

some intersections in Melb CBD I avoid altogether

if I know an intersection is a bad one I will avoid it

what constitutes a ‘bad’ intersection?

One where your onboard computer can’t read the traffic lights…

Eventually you might have enforced WIFI speed zones

buffy said:

stumpy_seahorse said:

CrazyNeutrino said:

some intersections in Melb CBD I avoid altogether

if I know an intersection is a bad one I will avoid it

what constitutes a ‘bad’ intersection?

One where your onboard computer can’t read the traffic lights…

in this case the computer is you

are there intersections where you cannot read the lights?

stumpy_seahorse said:

CrazyNeutrino said:

some intersections in Melb CBD I avoid altogether

if I know an intersection is a bad one I will avoid it

what constitutes a ‘bad’ intersection?

One where traffic signals are arranged so stupidly they create and add to congestion and delays rather than moderate them.

(.. I’m looking at you Roe Highway traffic signal engineers)

stumpy_seahorse said:

CrazyNeutrino said:

some intersections in Melb CBD I avoid altogether

if I know an intersection is a bad one I will avoid it

what constitutes a ‘bad’ intersection?

poor visibility, poor night lighting, too many ads, poor placement of signs, too many signs, too much visual information or not enough information, depends

google bad intersection design

have a look here

http://www.badintersections.com/

Bad Intersections

CrazyNeutrino said:

stumpy_seahorse said:

CrazyNeutrino said:

some intersections in Melb CBD I avoid altogether

if I know an intersection is a bad one I will avoid it

what constitutes a ‘bad’ intersection?

poor visibility, poor night lighting, too many ads, poor placement of signs, too many signs, too much visual information or not enough information, depends

google bad intersection design

any examples?

(admittedly, my driving experience in Melbourne’s CBD is limited, but I can’t think of an intersection in adelaide or hobart that fit those issues)

youtube > Bad Intersections

https://www.youtube.com/results?search_query=Bad+intersections

CrazyNeutrino said:

some intersections in Melb CBD I avoid altogether

if I know an intersection is a bad one I will avoid it

I will be driving in Melbourne in a few weeks. I am seriously considering just relying on public transport in the city and just using the car for the trip to Bendigo

Google > Bad Intersections > then click on Videos

https://www.google.com.au/search?q=bad+intersection+design&ie=utf-8&oe=utf-8&client=firefox-b&gfe_rd=cr&ei=iTzFV7jUGsTu8weN3bl4#q=bad+intersections&tbm=vid

Heaps of examples

>>too many ads, poor placement of signs, too many signs, too much visual information or not enough information<<

This is called visual confusion. Really. By the relevent engineers.

CrazyNeutrino said:

Google > Bad Intersections > then click on Videos

https://www.google.com.au/search?q=bad+intersection+design&ie=utf-8&oe=utf-8&client=firefox-b&gfe_rd=cr&ei=iTzFV7jUGsTu8weN3bl4#q=bad+intersections&tbm=vid

Heaps of examples

any australian?

you specified the melbourne CBD

As an out of left field box just make cars like dodgem cars and they all just bounce off each other.

AwesomeO said:

As an out of left field box just make cars like dodgem cars and they all just bounce off each other.

And watch the insurance companies run.

I think the best approach is to fallback on the minimum information to do the job reliably, which’d include it still being possible in the case of traffic lights failure. There’s a interesting one, how do you approach negotiating an intersection and other traffic if the traffic lights are dead.

transition said:

I think the best approach is to fallback on the minimum information to do the job reliably, which’d include it still being possible in the case of traffic lights failure. There’s a interesting one, how do you approach negotiating an intersection and other traffic if the traffic lights are dead.

standard road rules..

give way to the right

Did you know Google’s self-driving cars can’t handle 99% of roads in the US?

the article is two years old so is out of date

Many people have heard that Google’s autonomous cars can “drive anywhere a car can legally drive,” but it can’t drive in snow, heavy rains, see “unmapped” traffic lights or stop signs. In other words, Google’s self-driving cars can handle the “matrix” but it can’t navigate on 99% of the roads in the U.S.

Update sent by email from Kelly Mason, Global Communications & Public Affairs, Google:

The critical piece of misinformation that was reported was that the cars cannot detect “unmapped traffic lights or stop signs.” This is not true. The fact is that the car operates by combining mapped information with information collected in real time from the car’s sensors as it drives down the street. The combination of these two sources of information is what allows the car to drive safely — if there is an unmapped stop sign or traffic light, the sensors will identify it.

Yet another update!

On the backend, this article has caused a regular pooh storm and a flurry of emails. Google is unhappy. MIT is not particularly thrilled. Google wants you to know its cars would “identify” an umapped traffic light or stop sign, yet the fact is that MIT Lee Gomes has documented email proof that Chris Urmson, director of the car project, said a car could run a red light. Here is that quote:

Yes, if a traffic light were teleported into existence somewhere in the map, and the car didn’t know about it and there was no other traffic causing it to slow or stop, it could potentially run a red light. In practice this is very unlikely, as adding a traffic light takes time and construction and that this construction would have been detected by other cars, or from announcements made, and the traffic lights would be mapped when they came on line.

So Google’s self-driving cars might identify the light, but the real question – as pointed out courtesy of Gomes – the car may see it, but will it obey whatever the light or sign says?

self driving cars can park now

also see the tube video of the self driving Tesla P85D Supercar

stumpy_seahorse said:

transition said:

I think the best approach is to fallback on the minimum information to do the job reliably, which’d include it still being possible in the case of traffic lights failure. There’s a interesting one, how do you approach negotiating an intersection and other traffic if the traffic lights are dead.

standard road rules..

give way to the right

true, but you could give way right all day short of a break in traffic and opportunity, so you might nudge the nose out (something more than just a blinker announcement), then pull out and others follow close behind maintaining the stream.

Black boxes for P platers would be good

see too many weaving in in out of traffic

CrazyNeutrino said:

Black boxes for P platers would be good

see too many weaving in in out of traffic

see too many weaving in and out of traffic

CrazyNeutrino said:

CrazyNeutrino said:

Black boxes for P platers would be good

see too many weaving in in out of traffic

see too many weaving in and out of traffic

Compulsory web-cam?

PermeateFree said:

CrazyNeutrino said:

CrazyNeutrino said:

Black boxes for P platers would be good

see too many weaving in in out of traffic

see too many weaving in and out of traffic

Compulsory web-cam?

and compulsory GPS

and a system to detect weaving as a distinct car behavior

CrazyNeutrino said:

PermeateFree said:

CrazyNeutrino said:

see too many weaving in and out of traffic

Compulsory web-cam?

and compulsory GPS

and a system to detect weaving as a distinct car behavior

So weaving gets logged in the tamperproof black box

take that weavers

CrazyNeutrino said:

PermeateFree said:

CrazyNeutrino said:

see too many weaving in and out of traffic

Compulsory web-cam?

and compulsory GPS

and a system to detect weaving as a distinct car behavior

Ment to say Dash-cam, which should show most things a vehicle does.

PermeateFree said:

CrazyNeutrino said:

CrazyNeutrino said:

Black boxes for P platers would be good

see too many weaving in in out of traffic

see too many weaving in and out of traffic

Compulsory web-cam?

need to bring in compulsory drug and alcohol test before the ignition will work first..

stumpy_seahorse said:

PermeateFree said:

CrazyNeutrino said:

see too many weaving in and out of traffic

Compulsory web-cam?

need to bring in compulsory drug and alcohol test before the ignition will work first..

You having a go at Crazy?

PermeateFree said:

CrazyNeutrino said:

PermeateFree said:

Compulsory web-cam?

and compulsory GPS

and a system to detect weaving as a distinct car behavior

Ment to say Dash-cam, which should show most things a vehicle does.

yes and things like donuts, they have a specific car behavior

were talking suburb streets

not race tracks or tracks or roads allowing racing etc

stumpy_seahorse said:

PermeateFree said:

CrazyNeutrino said:

see too many weaving in and out of traffic

Compulsory web-cam?

need to bring in compulsory drug and alcohol test before the ignition will work first..

They already have alcohol ignition interlocks.

PermeateFree said:

stumpy_seahorse said:

PermeateFree said:

Compulsory web-cam?

need to bring in compulsory drug and alcohol test before the ignition will work first..

You having a go at Crazy?

nope.

if I was having a go at crazy, I’d want compulsory D&A testing before the forum let you log on…

CrazyNeutrino said:

stumpy_seahorse said:

PermeateFree said:

Compulsory web-cam?

need to bring in compulsory drug and alcohol test before the ignition will work first..

They already have alcohol ignition interlocks.

Ones for Cannabis will be coming out too

CrazyNeutrino said:

stumpy_seahorse said:

PermeateFree said:

Compulsory web-cam?

need to bring in compulsory drug and alcohol test before the ignition will work first..

They already have alcohol ignition interlocks.

should be mandatory, along with drug testing interlocks

stumpy_seahorse said:

CrazyNeutrino said:

stumpy_seahorse said:

need to bring in compulsory drug and alcohol test before the ignition will work first..

They already have alcohol ignition interlocks.

should be mandatory, along with drug testing interlocks

maybe it will be one day?

maybe it will only be for problematic drinkers, problematic drug addicts etc

the general public might erk at it

CrazyNeutrino said:

stumpy_seahorse said:

CrazyNeutrino said:

They already have alcohol ignition interlocks.

should be mandatory, along with drug testing interlocks

maybe it will be one day?

maybe it will only be for problematic drinkers, problematic drug addicts etc

the general public might erk at it

If my car was fitted with a cannabis ignition interlock

I would go by its judgement if I was fit or not to drive

but it has to take into account previous day activity as a low indicator starting point

I can drive unimpaired with cannabis in my system

by this I mean I have no visual impairment, no limb impairment, no imbalance impairment, no predictive impairment

in fact I feel better and have more attention

I might point out too that I have been driving for 35 years with no accident,

I haven’t been driving smoking pot for thirty five years, I generally drive when I’m not stoned, I don’t smoke and drive, and I don’t smoke then drive etc

If you know what I mean

But If I had a cannabis ignition interlock

Yes I would use it

CrazyNeutrino said:

CrazyNeutrino said:

stumpy_seahorse said:

should be mandatory, along with drug testing interlocks

maybe it will be one day?

maybe it will only be for problematic drinkers, problematic drug addicts etc

the general public might erk at it

If my car was fitted with a cannabis ignition interlock

I would go by its judgement if I was fit or not to drive

but it has to take into account previous day activity as a low indicator starting point

I can drive unimpaired with cannabis in my system

by this I mean I have no visual impairment, no limb impairment, no imbalance impairment, no predictive impairment

in fact I feel better and have more attention

I might point out too that I have been driving for 35 years with no accident,

I haven’t been driving smoking pot for thirty five years, I generally drive when I’m not stoned, I don’t smoke and drive, and I don’t smoke then drive etc

If you know what I mean

But If I had a cannabis ignition interlock

Yes I would use it

I think too the Government might look at whether or not to allow drivers to drive while on low dosages of prescribed medical cannabis

Many years ago one of the current affairs show did the experiment with alcohol, cannibis and some other drug of the day (possibly prescription mess that housewives use). The results were predictable within their ‘controlled’ driving experiment, but the interviews afterwards were interesting, with the cannibis driver saying that they thought they were ‘driving normally’ while they were really going quite slow with over exaggerated actions (taking a large circle to turn a corner etc). Safe, sure, but not ‘unimpaired’.

CrazyNeutrino said:

CrazyNeutrino said:

CrazyNeutrino said:

maybe it will be one day?

maybe it will only be for problematic drinkers, problematic drug addicts etc

the general public might erk at it

If my car was fitted with a cannabis ignition interlock

I would go by its judgement if I was fit or not to drive

but it has to take into account previous day activity as a low indicator starting point

I can drive unimpaired with cannabis in my system

by this I mean I have no visual impairment, no limb impairment, no imbalance impairment, no predictive impairment

in fact I feel better and have more attention

I might point out too that I have been driving for 35 years with no accident,

I haven’t been driving smoking pot for thirty five years, I generally drive when I’m not stoned, I don’t smoke and drive, and I don’t smoke then drive etc

If you know what I mean

But If I had a cannabis ignition interlock

Yes I would use it

I think too the Government might look at whether or not to allow drivers to drive while on low dosages of prescribed medical cannabis

Medical cannibis isn’t, generally, smoked. So the test would still stand.

Arts said:

Many years ago one of the current affairs show did the experiment with alcohol, cannibis and some other drug of the day (possibly prescription mess that housewives use). The results were predictable within their ‘controlled’ driving experiment, but the interviews afterwards were interesting, with the cannibis driver saying that they thought they were ‘driving normally’ while they were really going quite slow with over exaggerated actions (taking a large circle to turn a corner etc). Safe, sure, but not ‘unimpaired’.

impaired state is kind of the point of cannibis…,

stumpy_seahorse said:

Arts said:

Many years ago one of the current affairs show did the experiment with alcohol, cannabis and some other drug of the day (possibly prescription mess that housewives use). The results were predictable within their ‘controlled’ driving experiment, but the interviews afterwards were interesting, with the cannabis driver saying that they thought they were ‘driving normally’ while they were really going quite slow with over exaggerated actions (taking a large circle to turn a corner etc). Safe, sure, but not ‘unimpaired’.

impaired state is kind of the point of cannabis…,

indeed… I find it interesting when people who smoke it say they are unimpaired. I believe that they think they are because you do feel ‘relaxed but normal’ and can still seemingly function (unlike excessive alcohol) but it’s not the same as not smoking it

I sure as shit know when I am stoned mowing and I am most definitely impaired. It’s funny, you feel like you are going really fast and you focus a lot on being really precise. Anyone viewing would just think I am doing a normal mow but I feel,like I am at the edge of control.

Arts said:

Many years ago one of the current affairs show did the experiment with alcohol, cannibis and some other drug of the day (possibly prescription mess that housewives use). The results were predictable within their ‘controlled’ driving experiment, but the interviews afterwards were interesting, with the cannibis driver saying that they thought they were ‘driving normally’ while they were really going quite slow with over exaggerated actions (taking a large circle to turn a corner etc). Safe, sure, but not ‘unimpaired’.

What levels were they on?

I can imagine really high levels causing problems

by this I mean 20% TC or lower as a low indicator

this a normal for a joint or a few bongs

but say if the driver was way over 20% like 60% 80% 100% THC then that would be dangerous

people can pass out with high amounts of THC

you wouldnt get me in a car with someone who had just taken Cannabis Butter wat close to 100 % THC

but I would be ok with a regular joint user at 25% or lower THC

new uses and occasional users I would not get in the car with If they were driving stoned

because paranoia can be distracting, you dont get this with regular users

new users can get paranoid because their body is not used to the sensation, after all it it a new experience

like drinking alcohol for the first time then driving, dangerous

but with way less effects than alcohol but for first time users still dangerous because of paranoia or nausea

but Im not an expert that just my opinions

stumpy_seahorse said:

Arts said:

Many years ago one of the current affairs show did the experiment with alcohol, cannibis and some other drug of the day (possibly prescription mess that housewives use). The results were predictable within their ‘controlled’ driving experiment, but the interviews afterwards were interesting, with the cannibis driver saying that they thought they were ‘driving normally’ while they were really going quite slow with over exaggerated actions (taking a large circle to turn a corner etc). Safe, sure, but not ‘unimpaired’.

impaired state is kind of the point of cannibis…,

Not for me

I feel better

that is all

CrazyNeutrino said:

Arts said:

Many years ago one of the current affairs show did the experiment with alcohol, cannibis and some other drug of the day (possibly prescription mess that housewives use). The results were predictable within their ‘controlled’ driving experiment, but the interviews afterwards were interesting, with the cannibis driver saying that they thought they were ‘driving normally’ while they were really going quite slow with over exaggerated actions (taking a large circle to turn a corner etc). Safe, sure, but not ‘unimpaired’.

What levels were they on?

I can imagine really high levels causing problems

by this I mean 20% TC or lower as a low indicator

this a normal for a joint or a few bongs

but say if the driver was way over 20% like 60% 80% 100% THC then that would be dangerous

people can pass out with high amounts of THC

you wouldnt get me in a car with someone who had just taken Cannabis Butter wat close to 100 % THC

but I would be ok with a regular joint user at 25% or lower THC

new uses and occasional users I would not get in the car with If they were driving stoned

because paranoia can be distracting, you dont get this with regular users

new users can get paranoid because their body is not used to the sensation, after all it it a new experience

like drinking alcohol for the first time then driving, dangerous

but with way less effects than alcohol but for first time users still dangerous because of paranoia or nausea

but Im not an expert that just my opinions

I agree that the paranoia changes with more regular use… but I am no expert.

CrazyNeutrino said:

stumpy_seahorse said:

Arts said:

Many years ago one of the current affairs show did the experiment with alcohol, cannibis and some other drug of the day (possibly prescription mess that housewives use). The results were predictable within their ‘controlled’ driving experiment, but the interviews afterwards were interesting, with the cannibis driver saying that they thought they were ‘driving normally’ while they were really going quite slow with over exaggerated actions (taking a large circle to turn a corner etc). Safe, sure, but not ‘unimpaired’.

impaired state is kind of the point of cannibis…,

Not for me

I feel better

that is all

but if I take too much Cannabis I get randy as a Rabbit in China

work that one out

AwesomeO said:

I sure as shit know when I am stoned mowing and I am most definitely impaired. It’s funny, you feel like you are going really fast and you focus a lot on being really precise. Anyone viewing would just think I am doing a normal mow but I feel,like I am at the edge of control.

A low BGL for me is very similar to feeling stoned, and I start feeling it about 3.
they are now pushing the ‘BGL above 5 to drive’ campaign

Cannabis can exacerbate tiredness and sleepyness if you are already tired usually at night and at normal sleeptime ( but I would be in Bed anyway)

and this can become dangerous if driving because of Micro sleeps ( but I would be in Bed anyway)

so pull over and rest or go to sleep or go for a walk

stumpy_seahorse said:

AwesomeO said:

I sure as shit know when I am stoned mowing and I am most definitely impaired. It’s funny, you feel like you are going really fast and you focus a lot on being really precise. Anyone viewing would just think I am doing a normal mow but I feel,like I am at the edge of control.

A low BGL for me is very similar to feeling stoned, and I start feeling it about 3.
they are now pushing the ‘BGL above 5 to drive’ campaign

That’s what I’ve been told too, above 5.

stumpy_seahorse said:

PermeateFree said:

stumpy_seahorse said:

need to bring in compulsory drug and alcohol test before the ignition will work first..

You having a go at Crazy?

nope.

if I was having a go at crazy, I’d want compulsory D&A testing before the forum let you log on…

You wont get rid of me that way, as I don’t smoke or take drugs and drink very moderately. Therefore I would more likely be one of the few that could log on, although it would probably be less interesting.

You could have a smarter traffic light system spanning traffic lights, truck lights, car lights, bicycle lights motorcycle lights

Each traffic light for red yellow and green would be a specific color the driver less car could understand better and differentiate traffic lights from brake lights from vehicle lights thus reducing false positives even further

where each type of vehicle would have their own specific red hues for tail and braking which driverless cars could use for more road information

so:

trucks have specific red colors for tail and brake lights
cars have specific red colors for tail and brake lights
motorcycles have red specific colors for tail and brake lights
bicycles have specific red colors for tail and brake lights

similar for front lights slightly different white lights indicating type of vehicle

this gives another layer of information

probably in the future (if radio ID isn’t used) on traffic lights, the light itself can have or employ some sort signature, so that it can be quickly located in the near field by autonomous vehicle’ equipment.

like they shine a particular spectrum of light for starters. I think if a missile can navigate by stars during the day the technology to do this has to exist and be reliable.

I think with LED traffic lights it’s also possible to have them transmit information via LED (maybe using rolling code for security). IR LEDs are used all the time to transmit information. Radio’d be better, a small sensor in the lights that converts to radio would do (or connected back further) It radiates a signal that corresponds with light status (for an idea when it will transition too, the status and time to next change).

the radio’d probably need be spread spectrum like CDMA or whatever, to have immunity from noise and jamming (which’s what GPS has I reckon). + other encryption. Still jamming is possible.

CrazyNeutrino said:

You could have a smarter traffic light system spanning traffic lights, truck lights, car lights, bicycle lights motorcycle lights

Each traffic light for red yellow and green would be a specific color the driver less car could understand better and differentiate traffic lights from brake lights from vehicle lights thus reducing false positives even further

where each type of vehicle would have their own specific red hues for tail and braking which driverless cars could use for more road information

so:

trucks have specific red colors for tail and brake lights
cars have specific red colors for tail and brake lights
motorcycles have red specific colors for tail and brake lights
bicycles have specific red colors for tail and brake lights

similar for front lights slightly different white lights indicating type of vehicle

this gives another layer of information

another advantage of this is that computers can sense lights in daytime faster than humans

so emergency vehicles coming from behind could be detected earlier

You drove around in Melbourne without dropping in on me?

There are various difficulties in reading traffic lights. I can think of thee places in Melbourne where the traffic lights are particularly difficult to read. In one case the roads intersect at an angle just sufficiently skew that the right traffic light looks like the left, making the left turn arrow all but invisible. In a second case the lights at a dogleg crossing are confusing. In a third case the left green arrow is absent. It seems to be an unwritten law that an arrow light that is nonfunctional actually needs to be treated as a give way sign even when no give way sign is present.

In Melbourne, the whole concept of “lane” is also unreliable. When a road has no lanes marked it seems to be completely at the whim of the driver whether to treat it as a single lane or double lane. Overtaking on the left when no lanes are marked is a popular past time. It’s also sometimes left to the whim of individual drivers whether to treat a left-turning stretch of road as one lane or two. Dangerous.

An extra complication is that a few roads have “left turn on red”.
The hook turn in Melbourne where you turn right from the left lane can be confusing.

Also, a serious complication in Melbourne is that right arrow goes before straight ahead. This is wrong wrong wrong. Because it means two groups of cars turning right from the same street for every single cycle of the traffic lights.

There is also the complication that GPS maps are wrong. The GPS navigation frequently fails to identify the difference between a straight ahead road and a turn lane when the roads don’t intersect at right angles. It will also confuse a right turn with a left turn at some roundabouts. And gives roads wrong names.

CrazyNeutrino said:

CrazyNeutrino said:

You could have a smarter traffic light system spanning traffic lights, truck lights, car lights, bicycle lights motorcycle lights

Each traffic light for red yellow and green would be a specific color the driver less car could understand better and differentiate traffic lights from brake lights from vehicle lights thus reducing false positives even further

where each type of vehicle would have their own specific red hues for tail and braking which driverless cars could use for more road information

so:

trucks have specific red colors for tail and brake lights
cars have specific red colors for tail and brake lights
motorcycles have red specific colors for tail and brake lights
bicycles have specific red colors for tail and brake lights

similar for front lights slightly different white lights indicating type of vehicle

this gives another layer of information

another advantage of this is that computers can sense lights in daytime faster than humans

so emergency vehicles coming from behind could be detected earlier

so a line of traffic could alter more in tandem with the emergency vehicle, say fire or ambulance, police

Fix Melbourne? Yeah right!

mollwollfumble said:

You drove around in Melbourne without dropping in on me?

There are various difficulties in reading traffic lights. I can think of thee places in Melbourne where the traffic lights are particularly difficult to read. In one case the roads intersect at an angle just sufficiently skew that the right traffic light looks like the left, making the left turn arrow all but invisible. In a second case the lights at a dogleg crossing are confusing. In a third case the left green arrow is absent. It seems to be an unwritten law that an arrow light that is nonfunctional actually needs to be treated as a give way sign even when no give way sign is present.

In Melbourne, the whole concept of “lane” is also unreliable. When a road has no lanes marked it seems to be completely at the whim of the driver whether to treat it as a single lane or double lane. Overtaking on the left when no lanes are marked is a popular past time. It’s also sometimes left to the whim of individual drivers whether to treat a left-turning stretch of road as one lane or two. Dangerous.

An extra complication is that a few roads have “left turn on red”.
The hook turn in Melbourne where you turn right from the left lane can be confusing.

>>Also, a serious complication in Melbourne is that right arrow goes before straight ahead. This is wrong wrong wrong. Because it means two groups of cars turning right from the same street for every single cycle of the traffic lights.

how would that be any different to the straight ahead going before the turn green?

>>There is also the complication that GPS maps are wrong. The GPS navigation frequently fails to identify the difference between a straight ahead road and a turn lane when the roads don’t intersect at right angles. It will also confuse a right turn with a left turn at some roundabouts. And gives roads wrong names.

you need to get yourself a decent GPS.
every one that I have had clearly states the number of exit to take at a roundabout, ie..
“take 3rd exit at roundabout”

mollwollfumble said:

You drove around in Melbourne without dropping in on me?

There are various difficulties in reading traffic lights. I can think of thee places in Melbourne where the traffic lights are particularly difficult to read. In one case the roads intersect at an angle just sufficiently skew that the right traffic light looks like the left, making the left turn arrow all but invisible. In a second case the lights at a dogleg crossing are confusing. In a third case the left green arrow is absent. It seems to be an unwritten law that an arrow light that is nonfunctional actually needs to be treated as a give way sign even when no give way sign is present.

In Melbourne, the whole concept of “lane” is also unreliable. When a road has no lanes marked it seems to be completely at the whim of the driver whether to treat it as a single lane or double lane. Overtaking on the left when no lanes are marked is a popular past time. It’s also sometimes left to the whim of individual drivers whether to treat a left-turning stretch of road as one lane or two. Dangerous.

An extra complication is that a few roads have “left turn on red”.
The hook turn in Melbourne where you turn right from the left lane can be confusing.

Also, a serious complication in Melbourne is that right arrow goes before straight ahead. This is wrong wrong wrong. Because it means two groups of cars turning right from the same street for every single cycle of the traffic lights.

There is also the complication that GPS maps are wrong. The GPS navigation frequently fails to identify the difference between a straight ahead road and a turn lane when the roads don’t intersect at right angles. It will also confuse a right turn with a left turn at some roundabouts. And gives roads wrong names.

Yes there needs to be a faster error reporting system for GPS map errors

across all map providers

Lane Identification needs to be much improved

Lane lines and side lines, give way lines Stop lines various other road lines like double lines can be enhanced with Nano particles and specific colors driver less cars could use

CrazyNeutrino said:

mollwollfumble said:

You drove around in Melbourne without dropping in on me?

There are various difficulties in reading traffic lights. I can think of thee places in Melbourne where the traffic lights are particularly difficult to read. In one case the roads intersect at an angle just sufficiently skew that the right traffic light looks like the left, making the left turn arrow all but invisible. In a second case the lights at a dogleg crossing are confusing. In a third case the left green arrow is absent. It seems to be an unwritten law that an arrow light that is nonfunctional actually needs to be treated as a give way sign even when no give way sign is present.

In Melbourne, the whole concept of “lane” is also unreliable. When a road has no lanes marked it seems to be completely at the whim of the driver whether to treat it as a single lane or double lane. Overtaking on the left when no lanes are marked is a popular past time. It’s also sometimes left to the whim of individual drivers whether to treat a left-turning stretch of road as one lane or two. Dangerous.

An extra complication is that a few roads have “left turn on red”.
The hook turn in Melbourne where you turn right from the left lane can be confusing.

Also, a serious complication in Melbourne is that right arrow goes before straight ahead. This is wrong wrong wrong. Because it means two groups of cars turning right from the same street for every single cycle of the traffic lights.

There is also the complication that GPS maps are wrong. The GPS navigation frequently fails to identify the difference between a straight ahead road and a turn lane when the roads don’t intersect at right angles. It will also confuse a right turn with a left turn at some roundabouts. And gives roads wrong names.

Yes there needs to be a faster error reporting system for GPS map errors

across all map providers

Lane Identification needs to be much improved

Lane lines and side lines, give way lines Stop lines various other road lines like double lines can be enhanced with Nano particles and specific colors driver less cars could use

Id like to see road line paint that can last longer and not fade

CrazyNeutrino said:

mollwollfumble said:

You drove around in Melbourne without dropping in on me?

There are various difficulties in reading traffic lights. I can think of thee places in Melbourne where the traffic lights are particularly difficult to read. In one case the roads intersect at an angle just sufficiently skew that the right traffic light looks like the left, making the left turn arrow all but invisible. In a second case the lights at a dogleg crossing are confusing. In a third case the left green arrow is absent. It seems to be an unwritten law that an arrow light that is nonfunctional actually needs to be treated as a give way sign even when no give way sign is present.

In Melbourne, the whole concept of “lane” is also unreliable. When a road has no lanes marked it seems to be completely at the whim of the driver whether to treat it as a single lane or double lane. Overtaking on the left when no lanes are marked is a popular past time. It’s also sometimes left to the whim of individual drivers whether to treat a left-turning stretch of road as one lane or two. Dangerous.

An extra complication is that a few roads have “left turn on red”.
The hook turn in Melbourne where you turn right from the left lane can be confusing.

Also, a serious complication in Melbourne is that right arrow goes before straight ahead. This is wrong wrong wrong. Because it means two groups of cars turning right from the same street for every single cycle of the traffic lights.

There is also the complication that GPS maps are wrong. The GPS navigation frequently fails to identify the difference between a straight ahead road and a turn lane when the roads don’t intersect at right angles. It will also confuse a right turn with a left turn at some roundabouts. And gives roads wrong names.

Yes there needs to be a faster error reporting system for GPS map errors

across all map providers

Lane Identification needs to be much improved

Lane lines and side lines, give way lines Stop lines various other road lines like double lines can be enhanced with Nano particles and specific colors driver less cars could use

colors?..

what if it’s not an american car?

CrazyNeutrino said:

CrazyNeutrino said:

mollwollfumble said:

You drove around in Melbourne without dropping in on me?

There are various difficulties in reading traffic lights. I can think of thee places in Melbourne where the traffic lights are particularly difficult to read. In one case the roads intersect at an angle just sufficiently skew that the right traffic light looks like the left, making the left turn arrow all but invisible. In a second case the lights at a dogleg crossing are confusing. In a third case the left green arrow is absent. It seems to be an unwritten law that an arrow light that is nonfunctional actually needs to be treated as a give way sign even when no give way sign is present.

In Melbourne, the whole concept of “lane” is also unreliable. When a road has no lanes marked it seems to be completely at the whim of the driver whether to treat it as a single lane or double lane. Overtaking on the left when no lanes are marked is a popular past time. It’s also sometimes left to the whim of individual drivers whether to treat a left-turning stretch of road as one lane or two. Dangerous.

An extra complication is that a few roads have “left turn on red”.
The hook turn in Melbourne where you turn right from the left lane can be confusing.

Also, a serious complication in Melbourne is that right arrow goes before straight ahead. This is wrong wrong wrong. Because it means two groups of cars turning right from the same street for every single cycle of the traffic lights.

There is also the complication that GPS maps are wrong. The GPS navigation frequently fails to identify the difference between a straight ahead road and a turn lane when the roads don’t intersect at right angles. It will also confuse a right turn with a left turn at some roundabouts. And gives roads wrong names.

Yes there needs to be a faster error reporting system for GPS map errors

across all map providers

Lane Identification needs to be much improved

Lane lines and side lines, give way lines Stop lines various other road lines like double lines can be enhanced with Nano particles and specific colors driver less cars could use

Id like to see road line paint that can last longer and not fade

you want less jobs and higher unemployment?

stumpy_seahorse said:

CrazyNeutrino said:

mollwollfumble said:

You drove around in Melbourne without dropping in on me?

There are various difficulties in reading traffic lights. I can think of thee places in Melbourne where the traffic lights are particularly difficult to read. In one case the roads intersect at an angle just sufficiently skew that the right traffic light looks like the left, making the left turn arrow all but invisible. In a second case the lights at a dogleg crossing are confusing. In a third case the left green arrow is absent. It seems to be an unwritten law that an arrow light that is nonfunctional actually needs to be treated as a give way sign even when no give way sign is present.

In Melbourne, the whole concept of “lane” is also unreliable. When a road has no lanes marked it seems to be completely at the whim of the driver whether to treat it as a single lane or double lane. Overtaking on the left when no lanes are marked is a popular past time. It’s also sometimes left to the whim of individual drivers whether to treat a left-turning stretch of road as one lane or two. Dangerous.

An extra complication is that a few roads have “left turn on red”.
The hook turn in Melbourne where you turn right from the left lane can be confusing.

Also, a serious complication in Melbourne is that right arrow goes before straight ahead. This is wrong wrong wrong. Because it means two groups of cars turning right from the same street for every single cycle of the traffic lights.

There is also the complication that GPS maps are wrong. The GPS navigation frequently fails to identify the difference between a straight ahead road and a turn lane when the roads don’t intersect at right angles. It will also confuse a right turn with a left turn at some roundabouts. And gives roads wrong names.

Yes there needs to be a faster error reporting system for GPS map errors

across all map providers

Lane Identification needs to be much improved

Lane lines and side lines, give way lines Stop lines various other road lines like double lines can be enhanced with Nano particles and specific colors driver less cars could use

colors?..

what if it’s not an american car?

no im thinking more embedded nano particles in the road now

they could indicate on coming traffic signals, turns, roundabouts, stop signs giveaway signs etc to driverless cars

more information

>>You drove around in Melbourne without dropping in on me?

The hook turn in Melbourne where you turn right from the left lane can be confusing.<<

I had to spend all my time at Mum and Dad’s. Sorry.

The hook turn is brilliant, when done properly. Keeps a really good flow and doesn’t impede the trams. When I learnt to drive in 1977 we had to memorized which intersections (was it 6 or 8?) had hook turns, they were not marked. But they were the ones with trams in both streets. I think there are more of them now.

CrazyNeutrino said:

stumpy_seahorse said:

CrazyNeutrino said:

Yes there needs to be a faster error reporting system for GPS map errors

across all map providers

Lane Identification needs to be much improved

Lane lines and side lines, give way lines Stop lines various other road lines like double lines can be enhanced with Nano particles and specific colors driver less cars could use

colors?..

what if it’s not an american car?

no im thinking more embedded nano particles in the road now

they could indicate on coming traffic signals, turns, roundabouts, stop signs giveaway signs etc to driverless cars

more information

There is too much information now.

best I can come up with to find traffic lights and identify them (electronically using optical sensors/ cameras) in a hurry is to have an identifying code pulsed via LEDs (of course faster than people can notice with their eyes) Arranged so it can’t be imitated to mislead/mischief.

you’d look for the right colour, then look for the pulsed ID code in that area.