I really hate driving. Whether it’s short or long distance, I just think driving is a waste of time, especially when I get stressed out and use the f-word a hundred times while driving through the mazes in downtown Boston. I always thought I would definitely buy a car that can drive itself once there’s something available. There are news that Google is developing a driverless car, but I doubt that it will be commercially available to people like me in the near future.

Since developing a completely driverless car is so difficult, why not make a car that only needs a little bit of assistance to drive? You could imagine a car that can drive itself only in constrained environments such as in a freeway where the car merely needs to follow the white lanes while not bumping into the car in the front. We could designate the right most lane of freeways (which is the passing lane) to be the “autonomous driving lane”. Once the driver enters the lane, he could change his mode to autonomous driving. The driver can tell the car to just keep going until exit number X. When the car approaches exit X, it could have an alarm that tells the driver to take over driving again. There could be a shoulder next to the freeway so that the car can automatically pull itself over if the driver does not start driving again. Basically, the car can just take care of driving straight, following the white dotted lanes. Since the “autonomous lane” is the left-most lane, cars won’t be bothered by “human-driven cars” switching between lanes. That lane would be only for “autonomous driving”, possibly allowing the cars to communicate with each other to go at a fixed speed and maintain constant distance between cars. Since “autonomous driving” cars would be much more predictable than human-driven cars (computers do as they are programmed, humans don’t), they won’t need fancy algorithms that might be necessary for driverless driving in cities. This might help reduce traffic and maybe reduce car accidents since the autonomous driving cars would never tailgate nor go over the speed limit.

Compared to developing a completely driverless car, this technology would have a much bigger bang for the buck because the technology would be much easier to develop and drivers could be relieved from long-distance driving which is much more tiring than driving short distances in cities. Why bother developing the complicated technologies necessary for driving in cities when it’s not that grueling compared to all-day driving in freeways?

Maybe the main issue for both this technology and completely driverless cars is safety. Even if these technologies can reduce the number of car accidents, the auto companies, not the drivers, would be responsible for the accidents. If you remember how much damage Toyota had when 10-20 accidents were caused by malfunctioning Toyota Camrys, I am pretty sure customers will run away from autonomous driving cars once it leads to casualties. I bet that average customers won’t compare the possibility of an accident for the two cases and conclude that autonomous driving is safer. Instead, they will feel better to have control over their own fate (even thought that might increase the possibility of an accident) rather than rely on a computer to do the driving.

I wonder if there is going to be any way to overcome this problem. Autonomous driving  will become mainstream some day, but I bet it will never be perfectly accident-free. Any thoughts on how to solve this problem? Maybe customers would be willing to buy driverless cars if we can somehow prove that the accident rate is below a certain point? Who is going to pay for the insurance? Maybe car manufacturers can sell insurance with their autonomous cars.