Separate names with a comma.
Discussion in 'Autonomous Vehicles' started by daxz, Sep 2, 2016.
Nice youtube by CGP Grey - be sure to watch after "The End" (3:20 )
Saw the idea awhile back. This will never happen in your lifetime. Not a chance...
Autonomous vehicles would need to reach 100% saturation and manual driving would need to be banned. We'd live in a horrible Orwellian future where we are punished for thoughtcrime and animals start forming their own society.
Ya that will not happen. Keep dreaming city dweller.
I agree but I don't think it will ever happen. For one self driving cars would have to be perfect which they never will be. Second, I highly doubt you will be able to completely ban humans driving ever. The way we have it now, if someone messes up, that person and the one they hit are for the most part, the only ones involved. With this though, even if the rate of accidents is much lower, when it does happen, it will be really bad. Like 30-50+ cars. Another thing that a city dweller should know, what about pedestrians? How they going to cross if cars never stop? Build 4 bridge pathways at every intersection? I don't think so.
The ONLY place I could MAYBE see anything remotely close to this happening (where we ban human drivers and have self driving only) is in the very heart of cities during high traffic times. On off hours and outside the heart, you may see self driving cars, but you not going to completely ban humans driving.
I wrote most of the AI for an open-world game set in a large city several years ago, with computer-driven vehicles and pedestrians. They drove (and walked or ran) independently of each other aside from knowing distance, heading, and speed of nearby vehicles (and pedestrians), and were able to respect traffic signals and stop signs (and brake for errant pedestrians, like a player running across a street). With some logic for avoiding streets with heavy saturation (one vehicle per 20 feet or so) they were very good at avoiding traffic jams despite a very heavy saturation of vehicles. And never any accidents (of course, they knew where other vehicles were without having to use sensory systems to detect them).
But as soon as a human player started driving around, all hell broke loose. The biggest problem being that humans had a tendency to leave their vehicles in the middle of the road or sidewalk, necessitating support for u-turns and 3-point turns by the AI vehicles to overcome roads that became blocked by vehicles abandoned by human players. And players like to drive on the wrong side of the road at a very high speed, of course. But it still worked pretty well, for the most part.
Anyway, based on that experience I think a world where most vehicles are driven by AI could certainly work, especially with some universal communications protocol in the vehicles so they could know the position, heading, and velocity of every other nearby vehicle (including those being operated by humans). I can see this enabling very high speeds (with the vehicles tightly spaced) on freeways as well, with vehicles behind the leader of a "train" of vehicles managing their speed based on that of the vehicle immediately in front of them.
But it will be a long time until this is feasible, I think, if we need to rely on sensory systems to reliably detect anything other than stationary obstacles. On the other hand, I did that without much processing power and I had to simulate part of a city, not just a single vehicle, so maybe it will happen sooner than I imagine.
But we would have to completely ban humans driving at all times, I don't think that will ever happen. And everything I posted above on this topic from the horrible wreck when something does happen to pedestrians still stands. Especially the orwellian part....
I have a different view. Car [at the front of a line] crashes into something, maybe it is so sudden that the car behind runs into it, but by then all the other cars in the area will have communicated and come to a stop. Even if they can't communicate (some human drivers of old-tech are in the line too) then the 3rd car will have enough slack, from the car in front of it, coupled with 2nd car's "slack", to stop in time. Tesla is taking about 8.0 being able to see two-cars-in-front, not just the-car-in-front. Add into that car-to-car COMMs and things get even better.