The invention of the automobile has gone hand-in-hand with an unfortunate counterpart: accidents. This has been, mainly, due to driver error, rather than mechanical failure, so the objective for some time has been to eliminate this factor and make the daily commute less stressful. Popular culture brought us examples of cars that were capable of driving themselves, Knight Rider’s Knight Industries Three Thousand (KITT) for example, and it seems these once fairy tale visions will soon be realized.
The development of autonomous cars—and we’ll get to why this term is specifically used, rather than “driverless” vehicles—has been built on advances in safety that really started with seemingly innocuous features such as ABS. Here, an on-board computer senses heavy braking and intervenes before the wheels lock up. It’s therefore trying to circumvent human error, which is what autonomous cars do—just on a grander scale.
Volvo introduced its City Safety feature on their XC60 range some years back, which made use of lasers that could sense an impending accident. The technology assisted drivers by applying the brakes, up to 100% if required. The goal has been to use autonomous driving technology for those tedious commutes to work and back where frustration, tiredness, traffic jams and all those other aggravating factors come into play. This led to the development of FROG (Free Ranging On Grid technology); a car with an on-board computer loaded with maps. The vehicle knows where its starting location is and it checks its progress along the way by counting wheel revolutions, and its proximity to electronic beacons.
FROG is already being used for loading cargo as well as in factories, and the intention is to use it for public transportation. In this scenario, passengers go to a stop point and press a button; essentially calling the car as one would an elevator. You’d then get on and disembark at whichever location you’d selected as you would on a bus or train. Naturally, the downfall of the system is that the vehicles only go to predetermined places.
The autonomous cars of our future will make use of a range of technologies including radar, GPS, computer vision and would necessitate that the various systems of the car communicate with each other. Like the human brain, the aim is to take sensory information, interpret it and react in an appropriate manner. Some systems even allow for their maps to be updated as a result of changes to the environment. Once the technology is widespread enough, we could imagine a scenario where cars could “talk” to one another by sending warnings, to those further behind, of forthcoming obstacles.
Numerous car manufacturers have jumped on the autonomous vehicle bandwagon, but perhaps the most innovative leader in the field is not a maker of transportation, but search engine giant Google. The company has made use of retrofitted cars—usually a Toyota Prius or Lexus RX450H—loaded with LIDAR (Light Detection And Ranging), rangefinders, cameras and GPS sensors. Each car’s equipment costs in the region of $150,000 and there are usually 10 vehicles on the road at any given time, testing the systems thoroughly. The current expense of such technology is certainly a deterring factor for any prospective users, as polls have already indicated.
With time prices will decrease; in the future drivers may get some form of rebate or financial assistance from insurance companies. Offsetting the initial outlay are numerous benefits such as increased safety, fewer collisions, less congestion therefore increased road capacity, higher speed limits for autonomous cars, alleviation of parking problems—cars could drop us off and find parking far away and come collect us when required—and an increase in fuel efficiency.
Engineer Sebastian Thrun, the co-inventor of Google’s Street View and a director of the Stanford Artificial Intelligence Laboratory, started the Google autonomous car project. Extensive testing has taken place and the cars have driven on roads notorious for their choking traffic or dangerous bends, such as San Francisco’s Lombard Street.
Each car is accompanied by a driver, with an unblemished record, and a Google engineer in the passenger seat. In case of emergencies, the cars have an override system in place, which is very similar to those we find on modern cars with cruise control. Only two minor accidents have been recorded during the more than 482,000 kilometers of testing. Google claims that in both instances a person was driving, not the autonomous system and that the last instance was a bumper-bashing where the Google car got rear-ended.
The above begs consideration for two important points, firstly legislature concerning driverless cars and secondly, who is liable in the case of litigation. In all countries, the law presumes that a human adult is driving their vehicle. On March 1, 2012, the state of Nevada was the first to pass a law allowing for the operation of autonomous cars. In September, California and Florida followed suit.
Unsurprisingly, Nevada requires a person to be behind the wheel at all times to take over in emergency situations. License plates for future driverless cars in Nevada will be red and have an infinity symbol on the left-hand side. This of course limits the benefits of such vehicles as intoxicated drivers, minors or the unlicensed still couldn’t operate them.
In terms of the legal repercussions governing autonomous cars involved in accidents, some creative solutions have been put forward. One idea makes use of an example we’d never normally associate with driving: vaccines. The idea is that autonomous cars will be extremely safe but like a vaccine, where there is a tiny chance it could harm the patient, these vehicles could too. When a fatal accident occurs, according to this theory, the family would be compensated by a fund set up for this purpose.
Another school of thought is that the purchasers, not the manufacturers, of such cars take upon themselves the responsibility. This sounds reasonable until we consider that a software bug or mechanical failure could result in jail time or the driver being sued when in essence, he had no control over the situation. The more likely option is that the manufacturer would carry the blame, and costs incurred, should their product be found wanting. That’s a bitter pill to swallow as even if the percentage of defects is 1, this could still translate into extremely expensive lawsuits.
The last option is to spread the responsibility across all driverless car owners. They’d take out a special insurance package and the insurers would be responsible for determining whether the system, manufacturer or driver was at fault. This is still a complicated approach and difficult to implement but it would necessitate that car manufacturers would have to up their game which, in the long run, benefits everyone.
Jim Pisz, head of Toyota’s business strategy in North America, sums it up: “We feel that with autonomous driving, the driver should always be in control. It’s a critical element to our philosophy: autonomous does not mean driverless.”
Guess we won’t be calling KITT on our watch phone to save us from a sticky situation any time soon.