Advocacy & Safety - Self-Driving Cars

With experiments like the Opel Vectra and the DARPA competitions, cyclists have to look forward to the replacement of the distracted driver with self-driving cars. Not today, but it's coming.

http://screenrant.com/images/new-kitt2.jpg

What are your thoughts on this subject?

I think it will take 2-3 generations of self-drive to become acceptably reliable. It will come and go a few times, and more than a couple people will die because of the glitches.

When it finally arrives:

Road rage will be a thing of the past -- unless there is an override that isn't severely restricted, to emergencies and such. When people are 'being chauffeured', they will be a bit less agitated. Cell phone laws will be modified, if not outright stricken from the record. There will be more emphasis on creature comforts in the passenger cabin (hard to imagine more than NOW!), and an expansion of the already-in-place 'maintenance notiications' that send an e-mail to the owner when service or repair is needed.

A more realistic, and therefore unlikely, alternative to private robo-car ownership is the concept of mass, self-drive taxi service. Perhaps on a lease basis....

A more realistic, and therefore unlikely, alternative to private robo-car ownership is the concept of mass, self-drive taxi service. Perhaps on a lease basis....

They have something like this in Montreal http://www.communauto.com/comment_ENG.html There is a car pick-up point right next to where I live. However I prefer Montreal's Bixi bike rental system, but they havent yet extended their stations to my neighborhood.

The technology exists, and may soon be cost effective, for self-driving cars to avoid collision with lawfully operating cyclists. Tracking the same-direction motion of an upright warm body on the roadway ahead is easier than seeing the lines on the road through a construction zone in the rain and much easier than predicting the actions of children on a neighborhood street or even adult pedestrians downtown.

The big question in my mind is, will the level of caution required by the robot driver, in order to protect the deep-pocket car manufacturer from negligence lawsuits, require too slow a speed to suit the preferences of buyers?

That is, if a typical human driver would save time by passing at unsafe distance with oncoming traffic, but a robot driver would not, will car buyers want the robot driver?

I am reminded of Asimov's robot novels, where humans would confidently boss around robots, and often depended on the robots to rescue them, because the First Law. I wonder if cyclists might learn to do the same to robot drivers.

It will come slowly, and incrementally.

One step might be a vision system which provides gentle steering contra-pressure as the car drifts from the center of the lane, driver could still overcome and steer the car in whatever direction he wanted, but if left with hands off the wheel, the car would remain in the middle of the lane.

Another technology is a cruise control which also has a minimum following distance - as you approach another vehicle, the cruise would slack off to maintain a safe distance.

Perhaps I have too much faith in technology, but I fear the failures of these technologies far less than I fear the failures of human beings who think that they can control their automobiles.

By the time self-driving cars arrive, self-riding bikes will either have been invented, or the bicycle banned from public roadways.

Probably easier to ban bicycles than for the first generation of self-driving cars to be sensitive enough to track and avoid bicyclists.

Besides, banning a minority (cyclists) will be much easier, politically, than taking the control of cars away from the majority.

Interesting question.

My guess is that we will first see this in limited access situations, as a temporary way of dealing with the traffic generated by continued urban sprawl.

I think the current generation of Ford Rustang is a pretty fugly car to begin with, and slapping on a pile of body cladding (and not one, but two deck spoilers :rolleyes:) doesn't help.

Seriously, I think self-driving cars are are a couple rungs below flying cars on the pipe dream ladder. It's one thing to produce driver aids like radar/sonar to tell you the range to target is decreasing. To have all the systems integrated flawlessly so that the car "knows" where it and other objects are on the road under all conditions will prove to be extremely expensive and impractical.

But someone will try it, and it will be all fun and games till a gasoline tanker plows into a school after the truck's computer's copy of Windows 19 goes kaput while downloading a patch.

The technology exists, and may soon be cost effective, for self-driving cars to avoid collision with lawfully operating cyclists. Tracking the same-direction motion of an upright warm body on the roadway ahead is easier than seeing the lines on the road through a construction zone in the rain and much easier than predicting the actions of children on a neighborhood street or even adult pedestrians downtown.

The big question in my mind is, will the level of caution required by the robot driver, in order to protect the deep-pocket car manufacturer from negligence lawsuits, require too slow a speed to suit the preferences of buyers?

Why would speed matter if the occupants of the vehicle are occupied on the phone, or playing games or browsing on a computer.

Currently far too many drivers use speed to enhance the driving experience and make it more exciting.

If they don't even have to pay attention, and arrive at their designation on time, what does it really matter how fast they go?



That is, if a typical human driver would save time by passing at unsafe distance with oncoming traffic, but a robot driver would not, will car buyers want the robot driver?

I am reminded of Asimov's robot novels, where humans would confidently boss around robots, and often depended on the robots to rescue them, because the First Law. I wonder if cyclists might learn to do the same to robot drivers.

I suspect the "time saved" is really scant seconds. I know I have watched drivers that chose to weave in and out on the freeways while I maintain a steady speed and notice that we often arrive at the same exit (even 30 miles away) at practically the same time. (ultimately they may exit 10 cars in front of me, but I was close enough to see them exit) The fact is that a second here or a second there only adds up to anything significant over a very long drive.

The "delays" that motorist mention when following a cyclist are only 3-5 seconds typically. Traffic lights are more of a delay. Waiting for elevators is more of a delay.

The need for speed is more of an entertainment thing.

I think the current generation of Ford Rustang is a pretty fugly car to begin with, and slapping on a pile of body cladding (and not one, but two deck spoilers :rolleyes:) doesn't help.

Seriously, I think self-driving cars are are a couple rungs below flying cars on the pipe dream ladder. It's one thing to produce driver aids like radar/sonar to tell you the range to target is decreasing. To have all the systems integrated flawlessly so that the car "knows" where it and other objects are on the road under all conditions will prove to be extremely expensive and impractical.

But someone will try it, and it will be all fun and games till a gasoline tanker plows into a school after the truck's computer's copy of Windows 19 goes kaput while downloading a patch.

Actually Volvo has a self driving test car already on the streets. It is highly disguised and has a team of drivers to modify the software and verify the movements.

This is far closer to reality than flying cars.

Here is an interesting blog about what self driving cars may bring.
http://ideas.4brad.com/robocars-are-future

It will come slowly, and incrementally.


Another technology is a cruise control which also has a minimum following distance - as you approach another vehicle, the cruise would slack off to maintain a safe distance.



This cruise control already exists... it is called adaptive cruise control.

I think the current generation of Ford Rustang is a pretty fugly car to begin with, and slapping on a pile of body cladding (and not one, but two deck spoilers :rolleyes:) doesn't help.

Seriously, I think self-driving cars are are a couple rungs below flying cars on the pipe dream ladder. It's one thing to produce driver aids like radar/sonar to tell you the range to target is decreasing. To have all the systems integrated flawlessly so that the car "knows" where it and other objects are on the road under all conditions will prove to be extremely expensive and impractical.

But someone will try it, and it will be all fun and games till a gasoline tanker plows into a school after the truck's computer's copy of Windows 19 goes kaput while downloading a patch.

Without a new Pontiac to uglify....

I predict the availability of self-driving cars on freeways in 10-15 years, using some combination of GPS, LIDAR, RADAR and multispectral computer vision.

I suspect the systems will disable self driving outside of freeways due to the much more complex traffic and obstruction hazards on normal streets. I suspect they will use GPS to determine the type of roadway they are on, as well as the posted speed limit.

The question is, will car owners accept a robot driver that won't speed? I don't think the government will accept a robot driver that will speed.

Without a new Pontiac to uglify....

I predict the availability of self-driving cars on freeways in 10-15 years, using some combination of GPS, LIDAR, RADAR and multispectral computer vision.

I suspect the systems will disable self driving outside of freeways due to the much more complex traffic and obstruction hazards on normal streets. I suspect they will use GPS to determine the type of roadway they are on, as well as the posted speed limit.

The question is, will car owners accept a robot driver that won't speed? I don't think the government will accept a robot driver that will speed.

Probably the real question is when this technology will be adopted by Americans. The issues of liability and potential law suits tend to hold back developments in the US auto industry... unless mandated by law.

Probably the real question is when this technology will be adopted by Americans. The issues of liability and potential law suits tend to hold back developments in the US auto industry... unless mandated by law.

That's why I think it will be easier for the inventors to demonstrate a good safety record on freeways than on ordinary streets. On freeways, a driver is not faulted for colliding with traffic that comes unexpectedly at them from the side, but on ordinary streets, with blind pedestrians, children, etc. this happens all the time. It will be hard to develop a system that has the capacity to perform as well as a human driver can, which I think will be the benchmark mandated by the public and their government, despite the much lower bar set by what many human drivers do. A robot cannot communicate or judge the intentions of people and animals like a human can.

That's why I think it will be easier for the inventors to demonstrate a good safety record on freeways than on ordinary streets. On freeways, a driver is not faulted for colliding with traffic that comes unexpectedly at them from the side, but on ordinary streets, with blind pedestrians, children, etc. this happens all the time. It will be hard to develop a system that has the capacity to perform as well as a human driver can, which I think will be the benchmark mandated by the public and their government, despite the much lower bar set by what many human drivers do. A robot cannot communicate or judge the intentions of people and animals like a human can.

Kind of a strange irony that humans as drivers manage to kill 40,000 other humans a year and that is the metric.

"These cyclists slowed down my auto-driver so much until I dialed-down my cyclist passing distance setting to 3 inches."

If they don't even have to pay attention, and arrive at their designation on time, what does it really matter how fast they go?

This is why Amtrak and Greyhound are so wildly profitable.

With experiments like the Opel Vectra and the DARPA competitions, cyclists have to look forward to the replacement of the distracted driver with self-driving cars. Not today, but it's coming.


What are your thoughts on this subject?

As long as they follow the keep-it-simple concept and go step by step, learning and improving along the way, its a great idea. Automobiles and roadways are certainly less of a challenge to tackle than airplanes and flight paths, and the latter have been under 99% computer control for years.

roughstuff

I won't even own a car that has ABS or traction control. No way in hell I'm going to buy a car that drives itself.

I wonder if they will program the robot cars to honk and buzz bikes ;)

By the time this technology comes to fruition, we will have moved past the need for large numbers of single-occupancy vehicles because the fuel and room to run them will be on the downslope of availability. I don't know why we should spend a crapload of money to duplicate something that can be done cheaper, better, and more responsibly with mass transit.

I wonder if they will program the robot cars to honk and buzz bikes ;)

I will certainly reprogram my bicycle-mounted-collision-avoidance-transponder to emit a "wide load commercial vehicle" signal.

I won't even own a car that has ABS or traction control. No way in hell I'm going to buy a car that drives itself.

Won't you think about the car-hating cyclists who don't think people should be allowed to drive? :D

I won't even own a car that has ABS or traction control. No way in hell I'm going to buy a car that drives itself.


I wonder how many people said "I won't even own a car that has seatbelts or airbags" - fact is, you have very little choice in what 'features' they put on your car unless you do as I have done, and don't own a car at all.

As someone who sees drivers as being far more dangerous than cars, I would like to see the accellerator pedal removed from cars - I would trust a machine with watching out for my safety far more than my fellow man (or 90lb suburban housewife, drivin' in her S.U.V. )

Kind of a strange irony that humans as drivers manage to kill 40,000 other humans a year and that is the metric.

You're only thinking of the US. Worldwide it is on the order of a million fatalities per year, and most of the victims worldwide are pedestrians and bicyclists. A distant future where robot cars are programmed not to hit pedestrians and bicyclists might be worth considering.

As someone who sees drivers as being far more dangerous than cars, I would like to see the accellerator pedal removed from cars - I would trust a machine with watching out for my safety far more than my fellow man (or 90lb suburban housewife, drivin' in her S.U.V. )

All kinds of features could be incorporated as the software and knowlege base of such a system is improved and perfected. Once cars can sense one anothers presence, cars passing thru intersections would not be much more difficult than pedestrians at crossing sidewalks and paths...the speed and passage would be adjusted so that they flow continuously, rather than have the energy and time wasting stop and start we have now.

What makes anti-car luddites really shiver their timbers is they think a system like this might require that all roadway users be a part of the system, or be removed from the roadway entirely. Who knows. Maybe not! Maybe crazy urban assault cyclists would have a computer chip in their frame identifying them as such, and vehicles around them would make sure they received 3 feet, clear paths ahead to zoom to their destination between lanes, and prevent right and left hooks.

roughstuff

Worldwide it is on the order of a million fatalities per year, and most of the victims worldwide are pedestrians and bicyclists. A distant future where robot cars are programmed not to hit pedestrians and bicyclists might be worth considering.

Citation?

All I could find is this, which is for the US, and I don't think it covers bicycles. 11% of motor vehicle fatalities are pedestrians.
http://www.walkinginfo.org/pedsafe/crashstats.cfm

I can imagine pedestrian death rates would be higher in nations with.. slightly more chaotic driving.

Still, >50% sounds large.

This is why Amtrak and Greyhound are so wildly profitable.

Neither one delivers door to door service.

I recently tried doing a temporary commute of some 30+ miles using local train service... the last train left the station at 5:30PM. How freaking convenient.

Now imagine having a chauffeur... who happens to drive the speed limit and allows you to do whatever you want in the back seat... that is the model for self driving cars... not Amtrak or Greyhound.

You're only thinking of the US. Worldwide it is on the order of a million fatalities per year, and most of the victims worldwide are pedestrians and bicyclists. A distant future where robot cars are programmed not to hit pedestrians and bicyclists might be worth considering.

Exactly. :D

Neither one delivers door to door service.

I recently tried doing a temporary commute of some 30+ miles using local train service... the last train left the station at 5:30PM. How freaking convenient.

Now imagine having a chauffeur... who happens to drive the speed limit and allows you to do whatever you want in the back seat... that is the model for self driving cars... not Amtrak or Greyhound.

Buses and local trains are such an inconvenience that some communities are thinking of giving vouchers to use taxis and gypsy cabs, instead of subsidizing bus systems.

I am heading down to DC from WMass in a few days using the train...sure hope its better than your experinece Genec. Wouldn't want to fly or drive, though. Would have biked in better climate.

Anyway..for starters, I just wish 'smart cars' would do a few simple things...

(1) prevent drivers from going over the speed limit; and
(2) take a big step forward in avoiding accidents by preventing running red lights, rear end fender benders, etc. All that would be necessary for this is for the cars to 'talk' to the lights at an intersection, and with the car directly in front of them. Shouldn't be too hard.


roughstuff

Buses and local trains are such an inconvenience that some communities are thinking of giving vouchers to use taxis and gypsy cabs, instead of subsidizing bus systems.

I am heading down to DC from WMass in a few days using the train...sure hope its better than your experinece Genec. Wouldn't want to fly or drive, though. Would have biked in better climate.

Anyway..for starters, I just wish 'smart cars' would do a few simple things...

(1) prevent drivers from going over the speed limit; and
(2) take a big step forward in avoiding accidents by preventing running red lights, rear end fender benders, etc. All that would be necessary for this is for the cars to 'talk' to the lights at an intersection, and with the car directly in front of them. Shouldn't be too hard.


roughstuff

The sad thing with my train experience is that it was a beautiful setting, right along the coast... and if I combined bike with train it would have given me about 5-7 miles of bike commuting, and about 30 miles with the train. The train was very nice and had plenty of racks for bikes, but their schedule was hosed. If you work until 5:00, you need time to get to the station... I just could not believe the last train was at 5:30.

I ended up having to drive 34 miles each way for a month.

Now I have a new job 4.5 miles from home. 7.5 miles by the best streets to bike commute. And the best part is it takes over a half hour to drive as there are so many cars clogging the streets. I can actually bike it in 40 minutes... only because a bike lane exists. If there were no BL, it would be heck trying to fight the bumper to bumper traffic. Even at 8MPH, up the steep hills, I move far faster than motor traffic. :D

Citation?

All I could find is this, which is for the US, and I don't think it covers bicycles. 11% of motor vehicle fatalities are pedestrians.
http://www.walkinginfo.org/pedsafe/crashstats.cfm

I can imagine pedestrian death rates would be higher in nations with.. slightly more chaotic driving.

Still, >50% sounds large.

It's not just worse drivers and worse roads, but the fact that in much of the world, the majority of road users are pedestrians and cyclists, put in danger by a minority group of motorists. Also, many of them have a poor understanding of traffic safety because they are not trained drivers themselves.

In previous years it was over half; this year it seems to have dipped under half:

http://www.who.int/mediacentre/news/releases/2009/road_safety_report_20090615/en/index.html


Half of 1.27 million people who die in road traffic crashes every year are pedestrians, motorcyclists and cyclists, finds new WHO study.

I wonder how many people said "I won't even own a car that has seatbelts or airbags" - fact is, you have very little choice in what 'features' they put on your car unless you do as I have done, and don't own a car at all.

Well, to be fair, those devices have been found to increase survivability for crash victims by quite a bit. It makes sense that car manufacturers install these items on their cars. I would have no problem driving around without a seatbelt if I were the only driver on the road, but that is not the case. I have also removed airbags, ABS and cruise control from my car. Removing my airbag may be kinda silly, but it would only go off if I ran head-on into something. I plan on not doing that. I am still more protected than if I were on my bike.

ABS and traction control, however, remove control from the driver and are a crutch for those drivers who are under qualified or who lack common sense. Instead of paying attention and adjusting their driving to the road conditions, they rely on the car's systems to keep them on the road. "Oh, you don't know how to use your brakes correctly? That's ok, you can still drive because the brakes will sense your ham-feet stomping down and do the braking for you" or "It's cold, icy and snowy. You should know this and slow down, but instead this computer will try and keep you on the road. Go ahead and floor it." It's just asinine that cars and laws are designed around the population's worst drivers/citizens. It messes things up for the rest of us. As long as I can still drive a normal car without the laws or insurance companies forcing me to convert, I will be ok.

I'm getting off track. I'm done.

Well, to be fair, those devices have been found to increase survivability for crash victims by quite a bit. {snip snip snip**
I'm getting off track. I'm done.

Actually Glynis27, I don't think that you were very much off track. If drivers were well trained and responsible, there would be much less need for driving robots. The fact is that we can't rely on drivers being well trained, and that has led to things like ABS to make up for drivers who do not have the skill or training to operate their vehicles safely. Technology like ABS and traction control increase survivability and reduce crash rates for unskilled drivers. As a society we have decided that driving is a right, not a privilege - I know that legally it is not this way, but the reality is, it is very difficult to prove yourself such a bad driver that your 'privilege' is taken away. Given that, I don't think that it is that much of a stretch to allow robotic driving so that the right to drive can be truly universal.

Actually Glynis27, I don't think that you were very much off track. If drivers were well trained and responsible, there would be much less need for driving robots. The fact is that we can't rely on drivers being well trained, and that has led to things like ABS to make up for drivers who do not have the skill or training to operate their vehicles safely. Technology like ABS and traction control increase survivability and reduce crash rates for unskilled drivers. As a society we have decided that driving is a right, not a privilege - I know that legally it is not this way, but the reality is, it is very difficult to prove yourself such a bad driver that your 'privilege' is taken away. Given that, I don't think that it is that much of a stretch to allow robotic driving so that the right to drive can be truly universal.

I guess I can agree with this. I would rather have unskilled drivers being chauffeured around by computers instead of worrying about them crashing into me. It would make those few times I do drive even more enjoyable and less stressful. However, I do fear that laws or insurance would force everybody to have automatic driving cars, and I just won't go for that as long as I prove I am safe behind the wheel.

All kinds of features could be incorporated as the software and knowlege base of such a system is improved and perfected. Once cars can sense one anothers presence, cars passing thru intersections would not be much more difficult than pedestrians at crossing sidewalks and paths...the speed and passage would be adjusted so that they flow continuously, rather than have the energy and time wasting stop and start we have now.

What makes anti-car luddites really shiver their timbers is they think a system like this might require that all roadway users be a part of the system, or be removed from the roadway entirely. Who knows. Maybe not! Maybe crazy urban assault cyclists would have a computer chip in their frame identifying them as such, and vehicles around them would make sure they received 3 feet, clear paths ahead to zoom to their destination between lanes, and prevent right and left hooks.

roughstuff

Most likely a self-drive system would need to be able to detect other vehicles using some form of transponder, wouldn't be hard to mount a battery powered transponder on a bicycle,, might not even need to be visible. The self drive car would send out a radio wave, the transponder would receive that, and send a signal back. A transponder that is GPS enabled could encode it's position in the signal it sends out. The Self drive car knowing where the bicycle is, could easily make sure that it leaves the proper clearance when passing. When coming up to a stop sign the Self drive car would also detect the approaching bicycle, and act appropriately.

Where this could also be useful is traffic lights, the light could also make use of the transponder, detecting the bicycle and triggering a cycle change. The issue is that self drive cars are at least 20 years off, and coordinating adding transponders to all vehicles that could be on the road would probably take 10 years itself.

I guess I can agree with this. I would rather have unskilled drivers being chauffeured around by computers instead of worrying about them crashing into me. It would make those few times I do drive even more enjoyable and less stressful. However, I do fear that laws or insurance would force everybody to have automatic driving cars, and I just won't go for that as long as I prove I am safe behind the wheel.

Ultimately you may not have a choice... it could be that gasoline becomes scarce and the only fuel available is for self driving cars; your choice then would be either to bike or use a self driving car.

I will certainly reprogram my bicycle-mounted-collision-avoidance-transponder to emit a "wide load commercial vehicle" signal.

:thumb:

...ABS and traction control, however, remove control from the driver and are a crutch for those drivers who are under qualified or who lack common sense. Instead of paying attention and adjusting their driving to the road conditions, they rely on the car's systems to keep them on the road. ....

ABS makes adjustments much faster and much more reliably than almost ANY driver short of a NASCAR trained rider. For decades people told drivers to steer with the skid; but it is very counterintuitive. Nor can a driver assign the brake and power to different front wheels with anything RESEMBLING the accuracy and speed that the computer would do, if they can it at all.

Some cyclists resent auto technology because it increases the superiority of the automobiile as a mode of transportation. Wonderful agenda....if we could just make every other form of transportation less safe, , more people might ride bicycles. :)

roughstuff

http://www.youtube.com/watch?v=DRF_KaWzxq4&feature=PlayList&p=C98068F32328FA5D&playnext=1&playnext_from=PL&index=2

I can't remember where but I've seen self drive video tests. I'm imaging that the proximity and vectoring software regarding evasive/collision avoidance tech in these cars will simply factor in the cyclist as another logic process. Add to it the combination of heat signature/IR and body shape, then human avoidance priority settings would kick in. I do believe the main thrust of this technology will be highway driving to avoid pileups and help the commute along, and realize massive fuel savings enmass.

It wasn't the above video I saw, the one I saw was more about the logic and the sensors. I really do like the "learning" approach taken in that video ... very cool.

Most likely a self-drive system would need to be able to detect other vehicles using some form of transponder, wouldn't be hard to mount a battery powered transponder on a bicycle,, might not even need to be visible. The self drive car would send out a radio wave, the transponder would receive that, and send a signal back.
The system can't only depend on such active transponder devices since the car would need to react to a variety of road obstacles - fallen trees or branches, rock slides, deer and other animals, children, etc. that would not be transponder-equipped. But there are very good radar and sonar devices that can provide the data for detecting hazards in the car's path - including cyclists.

By the way one feature about self driving cars---if they indeed end up being safer, with fewer collisions etc---perhaps they could also be lighter and therefore more fuel efficient.

roughstuff

http://www.youtube.com/watch?v=DRF_KaWzxq4&feature=PlayList&p=C98068F32328FA5D&playnext=1&playnext_from=PL&index=2

Very interesting video clip. Why is that the new cars don't have the ability to communicate with each other?

Given that most if not all newer cars have GPS' built into them. The ability for cars to be able to communicate with each other they would be able to know where each car is on the road.

For bicycles and horse and buggies and other similar vehicles, the stand alone GPS' could also be programmed to communicate with the GPS' in the cars. So that their road position could be known.

And through having the cars communicating with each other sharing GPS data if the following/passing car got to close it could be programmed to slow down to a safer speed.

Also all of those ubiquitous road reflectors could be setup to communicate with the car. So that if it's raining the car could be slowed down, likewise if the car is approaching a blind curve.

In that way the driver would still have control of the car, but in certain circumstances the car can take limited control.

Also with with the cars GPS' communicating with each other emergency vehicles would also be able to override cars and slow them down and direct them towards the side of the road.

Very interesting video clip. ........Also with with the cars GPS' communicating with each other emergency vehicles would also be able to override cars and slow them down and direct them towards the side of the road.



Goodness some very interesting applications. Is there an app for that? In any case, it shows an excellent example of how self driving cars can drastically improve performance on the roadways. To keep the system sane and minimize 'learning curve errors,' I hope we introduce these new applications with some patience.

roughstuff

Very interesting video clip. Why is that the new cars don't have the ability to communicate with each other?

Given that most if not all newer cars have GPS' built into them. The ability for cars to be able to communicate with each other they would be able to know where each car is on the road.

For bicycles and horse and buggies and other similar vehicles, the stand alone GPS' could also be programmed to communicate with the GPS' in the cars. So that their road position could be known.

And through having the cars communicating with each other sharing GPS data if the following/passing car got to close it could be programmed to slow down to a safer speed.

Also all of those ubiquitous road reflectors could be setup to communicate with the car. So that if it's raining the car could be slowed down, likewise if the car is approaching a blind curve.

In that way the driver would still have control of the car, but in certain circumstances the car can take limited control.

Also with with the cars GPS' communicating with each other emergency vehicles would also be able to override cars and slow them down and direct them towards the side of the road.

Our current GPS system while quite accurate using the right algorithms, is somewhat dependent on accurate maps, which we don't yet have. Of course over time, the demand for better data could result in the accuracy we need.

Autonomous cars cannot use GPS as the primary navigator. Any communication lost between car/car/sat could be disastrous and hackable. All systems will need to be self contained within the vehicle to be viable. A "grid" of cars being controlled by a single source could be terrible news.

Autonomous cars cannot use GPS as the primary navigator. Any communication lost between car/car/sat could be disastrous and hackable. All systems will need to be self contained within the vehicle to be viable. A "grid" of cars being controlled by a single source could be terrible news.

true that. the Darpa challenges allowed GPS use only for coarse course correction.

http://www.technologyreview.com/read_article.aspx?id=18908
The DARPA challenge?

DARPA (the Defense Advanced Research Projects Agency) test the car and determine if it proceeds to the next round in the Urban Challenge.
LEARN ABOUT DARPA

DARPA is the research and development arm of the U.S. Department of Defense. DARPA's mission is to maintain the technological superiority of the U.S. military and prevent technological surprise from threatening our national security.

DARPA funds unique and innovative research through the private sector, academic and other non-profit organizations as well as government labs.

DARPA research runs the gamut from conducting scientific investigations in a laboratory, to building full-scale prototypes of military systems. We fund research in biology, medicine, computer science, chemistry, physics, engineering, mathematics, material sciences, social sciences, neuroscience, and more.

DARPA research projects evolve into military use through technology transition.

This is NOT who you want to control your autonomous car technology. You think you're making it safer for everyone on the roads and then your research is suddenly installed into autonomous drones in the air and on the ground. Self guided car bombs, hunter/killer air drones (which we already have), ah no more DARPA entries please...back to lab geeks we're selling the next one to Mercedes. :)

Autonomous cars cannot use GPS as the primary navigator. Any communication lost between car/car/sat could be disastrous and hackable. All systems will need to be self contained within the vehicle to be viable. A "grid" of cars being controlled by a single source could be terrible news.


Correct. IT also violates the KISS principle. WE need to walk slowly before we run, and learn along the way. The first task is to allow cars to 'talk to one another ' and avoid accidents; talk to traffic lights and and coordinate movement; etc. Zillions of cars as part of a satellite network would be considerable further down the line.

roughstuff

http://www.technologyreview.com/read_article.aspx?id=18908
The DARPA challenge?

DARPA (the Defense Advanced Research Projects Agency) test the car and determine if it proceeds to the next round in the Urban Challenge.
LEARN ABOUT DARPA

DARPA is the research and development arm of the U.S. Department of Defense. DARPA's mission is to maintain the technological superiority of the U.S. military and prevent technological surprise from threatening our national security.

DARPA funds unique and innovative research through the private sector, academic and other non-profit organizations as well as government labs.

DARPA research runs the gamut from conducting scientific investigations in a laboratory, to building full-scale prototypes of military systems. We fund research in biology, medicine, computer science, chemistry, physics, engineering, mathematics, material sciences, social sciences, neuroscience, and more.

DARPA research projects evolve into military use through technology transition.

This is NOT who you want to control your autonomous car technology. You think you're making it safer for everyone on the roads and then your research is suddenly installed into autonomous drones in the air and on the ground. Self guided car bombs, hunter/killer air drones (which we already have), ah no more DARPA entries please...back to lab geeks we're selling the next one to Mercedes. :)

Uh, no, we don't want DARPA controlling our cars, but like many other technological advances from air travel to computers to cell phones, a government/military initiative was the first step that allowed those advances to be made available to the public. Even our interstate freeways and GPS systems all have military history.

And yes, the DARPA challenge has already succeeded in producing self drive vehicles... although I understand that DARPA also chose a funding path for one company based on their military-connection basis rather than their ability to meet the challenge.

Volvo and Audi are already working on self drive technology.

http://www.self-drivingcar.com/blogs.php
http://www.motorward.com/2009/10/self-driving-audi-tt-goes-rallying/
http://www.motorauthority.com/blog/1037007_sartre-initiative-working-on-self-driving-cars

Even VW has thrown it's hat into the ring.
http://www.theregister.co.uk/2006/07/04/self_driving_golf/

Top Gear in a self drive BMW on a track... at race speed.
http://videos.streetfire.net/video/126-Top-Gear-Self-Driving_180380.htm

So I certainly would not worry about DARPA being the "the one in control." :D