Self-Driving Vehicles
03-07-16, 09:35 AM
#256
Lexus Champion
From a known position, knowing the vehicle's heading (horizontal and vertical direction), and speed and time travelled (from which distance travelled can be calculated), it is possible to calculate the current position of the vehicle. Airliners travel from radio beacon to radio beacon, and since the exact location of each beacon is known, it is possible to use a radio beacon as a starting point for dead reckoning.
When I worked on designing automated train controls, we did the same thing: Along the tracks are beacons that are spaced a known distance apart. To determine location between beacons, the vehicle speed and time from the last beacon would be used to calculate distane travelled. We, of course, had to take into account wheel diameter and were worried about wearing down (resulting in smaller diameter) of the steel wheels, which would give the effect of faster speed and greater distance travelled.
Automated, connected cars can and should be using dead reckoning for navigation.
Our modern smartphones can also calculate position when it does not have line-of-sight to GPS satellites, by using cell towers: Knowing the location of at least 3 nearby cell towers and knowing the distance to each, it is possible to triangulate the exact position between them. That is also how GPS works: knowing the distance (by measuring exact time) to each GPS satellite, exact location (latitude, longitude and height / altitude) can be triangulated. Our connected cars could do this also.
Toyota (and other automakers and mapmakers) is also generating databases of location highlights -- significant buildings, towers, etc. -- that the car would recognize, just as we recognize such artifacts by sight as we travel. This can also be done as we travel: As the car travels along, the cameras all around the car would be taking pictures and sending these to a central database, much like Google Streetview, but in real time.
All of this technology should be used together, not only as backup but as a secondary, confirming source of location.
03-10-16, 02:23 AM
#257
In pre-GPS days, navigation was by dead reckoning; aircraft use it and automated, driverless trains use it.
From a known position, knowing the vehicle's heading (horizontal and vertical direction), and speed and time travelled (from which distance travelled can be calculated), it is possible to calculate the current position of the vehicle. Airliners travel from radio beacon to radio beacon, and since the exact location of each beacon is known, it is possible to use a radio beacon as a starting point for dead reckoning.
When I worked on designing automated train controls, we did the same thing: Along the tracks are beacons that are spaced a known distance apart. To determine location between beacons, the vehicle speed and time from the last beacon would be used to calculate distane travelled. We, of course, had to take into account wheel diameter and were worried about wearing down (resulting in smaller diameter) of the steel wheels, which would give the effect of faster speed and greater distance travelled.
Automated, connected cars can and should be using dead reckoning for navigation.
Our modern smartphones can also calculate position when it does not have line-of-sight to GPS satellites, by using cell towers: Knowing the location of at least 3 nearby cell towers and knowing the distance to each, it is possible to triangulate the exact position between them. That is also how GPS works: knowing the distance (by measuring exact time) to each GPS satellite, exact location (latitude, longitude and height / altitude) can be triangulated. Our connected cars could do this also.
Toyota (and other automakers and mapmakers) is also generating databases of location highlights -- significant buildings, towers, etc. -- that the car would recognize, just as we recognize such artifacts by sight as we travel. This can also be done as we travel: As the car travels along, the cameras all around the car would be taking pictures and sending these to a central database, much like Google Streetview, but in real time.
All of this technology should be used together, not only as backup but as a secondary, confirming source of location.
From a known position, knowing the vehicle's heading (horizontal and vertical direction), and speed and time travelled (from which distance travelled can be calculated), it is possible to calculate the current position of the vehicle. Airliners travel from radio beacon to radio beacon, and since the exact location of each beacon is known, it is possible to use a radio beacon as a starting point for dead reckoning.
When I worked on designing automated train controls, we did the same thing: Along the tracks are beacons that are spaced a known distance apart. To determine location between beacons, the vehicle speed and time from the last beacon would be used to calculate distane travelled. We, of course, had to take into account wheel diameter and were worried about wearing down (resulting in smaller diameter) of the steel wheels, which would give the effect of faster speed and greater distance travelled.
Automated, connected cars can and should be using dead reckoning for navigation.
Our modern smartphones can also calculate position when it does not have line-of-sight to GPS satellites, by using cell towers: Knowing the location of at least 3 nearby cell towers and knowing the distance to each, it is possible to triangulate the exact position between them. That is also how GPS works: knowing the distance (by measuring exact time) to each GPS satellite, exact location (latitude, longitude and height / altitude) can be triangulated. Our connected cars could do this also.
Toyota (and other automakers and mapmakers) is also generating databases of location highlights -- significant buildings, towers, etc. -- that the car would recognize, just as we recognize such artifacts by sight as we travel. This can also be done as we travel: As the car travels along, the cameras all around the car would be taking pictures and sending these to a central database, much like Google Streetview, but in real time.
All of this technology should be used together, not only as backup but as a secondary, confirming source of location.
Still doesn't answer the "human equation" of some hobo jumping out in the street, some idiot running a red light, the city bus "not giving right away", etc. The human brain at this point has such a huge advantage over a computer in processing "unknown or irregular" variables like those mentioned that it seems like an almost insurmountable hurdle to have a computer know how to react in the randomness that is our current street system.
I predict that the driver less car, on our current road system, with our current legal system, will not become viable until somebody like Commander Data becomes viable. Then we enter a whole new problem, that they couldn't figure out in the 24th century, is artificial intelligence classified as a life form? Is it a sentient being? What sort of rights do we grant the those that are the result of our own genius? I'm 32 now, and if I live to be 90 or 100 years old, I can see these sort of things being argued about in the Supreme Court. I feel at the rate we are progressing, Mr Data might happen in my lifetime.
Last edited by Aron9000; 03-10-16 at 02:31 AM.
03-10-16, 12:24 PM
#258
would like to look at the societal impact of driverless cars. specifically, jobs. the impact could (will) be gigantic.
school buses? no drivers.
taxis? no drivers.
truck transportation? no drivers.
trains (should have happened long ago)? no drivers.
warren buffett thinks the entire vehicle insurance industry will collapse which would have a huge economic impact.
personal cars? really little point for many if it would be MUCH cheaper to use a driverless uber type service. i can even imagine ONE PASSENGER driverless vehicles which would be very economical.
if most people end up not driving there may come a time when driving schools go out of business too, as well as revenue/administration from licensing, registration, etc.
we're really looking at a transforming change.
school buses? no drivers.
taxis? no drivers.
truck transportation? no drivers.
trains (should have happened long ago)? no drivers.
warren buffett thinks the entire vehicle insurance industry will collapse which would have a huge economic impact.
personal cars? really little point for many if it would be MUCH cheaper to use a driverless uber type service. i can even imagine ONE PASSENGER driverless vehicles which would be very economical.
if most people end up not driving there may come a time when driving schools go out of business too, as well as revenue/administration from licensing, registration, etc.
we're really looking at a transforming change.
03-11-16, 05:48 AM
#259
03-11-16, 07:08 AM
#260
GM buys self-driving car kit startup Cruise, plans to use tech to make driverless cars
http://techcrunch.com/2016/03/11/gm-...iverless-cars/
http://techcrunch.com/2016/03/11/gm-...iverless-cars/
03-11-16, 07:09 AM
#261
What happens if a special needs child falls asleep on the bus after dropping all the other kids off and gets stuck on the bus? Then what? Will a camera be able to stop bullying or confrontations on the bus?
03-11-16, 07:11 AM
#262
Lexus Fanatic
iTrader: (20)
Still doesn't answer the "human equation" of some hobo jumping out in the street, some idiot running a red light, the city bus "not giving right away", etc. The human brain at this point has such a huge advantage over a computer in processing "unknown or irregular" variables like those mentioned that it seems like an almost insurmountable hurdle to have a computer know how to react in the randomness that is our current street system.
I predict that the driver less car, on our current road system, with our current legal system, will not become viable until somebody like Commander Data becomes viable.
it could also be that driverless vehicles enter the commercial space long before the personal space, such as trucking, delivery vans, taxis, etc.
i think there will be some available for sale within 5 years.
03-11-16, 08:33 AM
#264
Since you have identified the problem, why don't you also propose choice of solutions to the group, rather than just one solution that you believe is the only solution that will work.
03-11-16, 08:37 AM
#265
..
human has huge advantage? lol. the computer isn't texting while driving, playing with the stereo, chatting to passengers, oggling a hot person in another car. it doesn't have bad eye sight, doesn't sneeze or cough, doesn't fumble around looking for something dropped in the seat crack, doesn't mess with a touch screen or other controllers, doesn't get road rage, isn't doing make-up, need i go on?
....
i think there will be some available for sale within 5 years.
human has huge advantage? lol. the computer isn't texting while driving, playing with the stereo, chatting to passengers, oggling a hot person in another car. it doesn't have bad eye sight, doesn't sneeze or cough, doesn't fumble around looking for something dropped in the seat crack, doesn't mess with a touch screen or other controllers, doesn't get road rage, isn't doing make-up, need i go on?
....
i think there will be some available for sale within 5 years.
All these humans just trying to ensure they don't get pushed over by machines are raising on-the-fringe doubts to shoot the idea down. Isn't it pathetic?
The human mind is superior? less than 1% of humans have superior minds. the rest are just riding on their laurels and imagining themselves to be that 1%
03-11-16, 08:48 AM
#266
I'm just saying an autonomous school bus will not replace a human driver/supervisor. There will always need to be an adult on that bus. Yes, the bus will be autonomous but you still need to pay someone to supervise the children.
03-11-16, 08:50 AM
#267
Just read this and it really puts things into perspective.
http://www.wired.com/2016/03/sadness...es-ai-play-go/
http://www.wired.com/2016/03/sadness...es-ai-play-go/
03-11-16, 10:04 AM
#268
The problem is how to take care of special needs children on an autonomous bus.
You have only one solution, that you have predetermined as the best and only solution.
03-11-16, 10:16 AM
#269
I've proposed a solution by having an adult supervisor on the autonomous bus. Are you trying to say there is a better solution?
03-11-16, 10:20 AM
#270