With all the talk about self-driving cars, I’ve never seen this issue discussed.
But it seems critical.
Some autonomous cars still have steering wheels, but the Google car very proudly does not.
There’s a lot of talk about how people are unreliable, fall asleep, etc, and can’t be trusted to safely take over the controls of a self-driving car. But there are situations where I simply can’t imagine how it is imposssible NOT to give control to the human.
A typical shopping mall has a huge parking lot, many entrances to the main builiding, and many other separate buildings and restaurants located inside the area of the asphalt. After the self-driving car turns off the main road into the parking area, you may have to drive another half mile, and your final stopping point will be different on each trip.
How do the self-driving cars which are currently on the road handle this situation?
And what are the engineers at Ford and Toyota planning for the mass-produced cars which they seem so sure will be on the market very soon?
Maybe leave the steering wheel in place, but allow the driver to take over only after the car brings itself to a stop? Maybe voice control to give instructions to the car?
Assuming a fully-autonomous car, the simpler solution would seem to be that you instruct it on your actual intended destination (i.e. which shop or restaurant you want to visit), and it knows where that is, and tries to park you as close as it can.
(It then sells the information about your shopping choices to the highest bidder)
ETA: of course Son of a Rich’s answer is much better. Goes to show that we have some new thinking to do about these technologies.
I think that is an interesting question. When I go to a parking lot I probably use a simple algorithm in that I see how busy the parking lot is, and how close to the store entrance I am likely to find a spot. I often start with the row I would most like a spot in, and then circle out. The busier the lot, the less picky I will be. I imagine a car would follow a similar routine, and I could tell it which entrance I would like to be closest to. If cars eventually communicate with each other directly, maybe the car would use an internal map of the lot, and drive directly to an open spot.
I don’t think the steering wheel-less car is really meant for consumers. It’s meant to be a taxi/Uber/etc. You call up an AI car, it picks you up, drops you off, and then goes on it’s way.
I don’t see how this is critically difficult. Let’s say you want to go to a specific store in the mall – you ask the car that your destination is that specific store in the mall, and the software will ideally calculate and drop you off at the closest entrance, and (after dropping you off) find the closest available parking spot. If on the other hand the “mall” is your destination, without any specific store, the software would presumably just drop you off at the entrance closest to the main road. It seems like that’s just how a human driver would deal with the situation.
But like Mangetout says, an autonomous car is going to fundamentally change the entire idea of car ownership that the question may be moot. Why does it matter where the car parks if it makes more sense for it to drop you off at the entrance, and later come pick you up when you’re ready to go home? More importantly, why even have the car sitting at the parking lot? Think of how much money car ownership costs in terms of the actual time we spend using it. In an average 24 hour day, my car sits in a parking space/garage more than 23 hours – that means I only use it less than 5% of the time. Multiply that by the millions of people who own cars, and the amount of wasted potential is staggering. Autonomous cars that can drop us off and pick us up on request will likely change that type of thinking, and we’re probably going to see a completely new type of car “ownership” experience that’s closer to car sharing where the car can be used by another person instead of sitting in a parking space.
I’ve seen car parks which had a sensor on every bay - so the car park knows which spaces are empty or full and displays signs for drivers to decide whether they want to branch off and park, or keep going for more spaces ahead - it seems like it wouldn’t be very difficult for a self-driving car to be networked to such a system and to already know where the spaces are (perhaps also actually reserving the space that they are intending to use).
i have seen documentaries relating to this … and have no problem embracing this environment … long as there is an option/request for a non-smoker vehicle.
In other words, because the founders did not write an amendment guaranteeing the right to bear vehicles, they’re taking them away from us. Without a whimper of protest from the car nuts.
It’s amazing how many questions there are along the lines of “how would an automated car handle this?”, where the answer is “the same way a human would, or better”.
Most likely, via their sonar or radar return, but they could also use the fact that a grocery bag looks nothing like a child, the same as a human would.
Now you tell me: Given a grocery bag just sitting in the road, not blowing, how would a human driver differentiate whether it has something big and solid in it, or if it’s just empty?
I have visions of a downtown Manhattan full of empty cars just circling the blocks in holding patterns waiting for their master. Gas or electricity is cheaper than downtown parking.
I would also suggest maybe the car would have in its GPS database which mall doors are double doors, so it could push its way into the mall and drop you in front of the requested store.
I guess proper marking of handicapped spots will be a necessity too.
More likely, I expect the mall to set up a dense parking area for self-drive, much like those downtown lots where you leave the keys. They park bumper to bumper with no room after dropping you at the door. When you push the remote to call your car, it communicates to the other cars, and they back out in a procession like a marching band until it can get free. The valet version includes a charger in each spot… and a human to plug them in.
That’s one easy. There are numerous very good algorithms for identifying human shaped objects. Histogram of Oriented Gradients comes to mind immediately (just because it is so well known) but there are others.
Self driving cars won’t replace your personal car. They’ll replace taxis, which will replace your personal car. Do taxis drop you off in enormous parking lots? No. Part of the benefit of self driving cars is that we get to eliminate parking lots!
