Last year, we wrote about Google developing technology that would allow cars to drive themselves. At that time, we questioned the legal impact of such a vehicle, both in the legislative and liability arenas. Now, we must consider the actual effectiveness of these systems in preventing car accidents and motorcycle accidents.
A number of automobile manufacturers are looking to get into the automated vehicle game. Google has been testing Toyota Priuses and Audi TTs with the automated systems and have logged over 140,000 miles. While perhaps not as far advanced as Google’s automated cars, Volkswagen has a Temporary Autopilot system (TAP) for use on highways. TAP maintains a safe distance between vehicles, reduces speed as necessary in a curve and also keeps the car within lane markers. It does all this while also adhering to passing rules and speed limits. But a human driver is required to constantly monitor the TAP system to react to dangerous situations as they arise. TAP seems to be aimed at preventing accidents due to driver inattentiveness or distractions. As with the VW system, the Google cars must always have a human driver ready to take over if required. But, what’s to prevent a bored driver who is surfing the internet to ignore safety warnings issued by the system?
There are potential safety advantages to a driverless car. People who drive while intoxicated would have a computer available to take over for them on the drive home from a party. People who cannot extricate themselves from their internet addiction would be doing everyone a favor by allowing a computer to take over driving rather than subjecting the rest of us to their selfish distracted driving blunders.
How well the systems sense different things on the road must be carefully evaluated. Automated cars may be quite adept at maintaining a safe distance from 18 wheelers barreling down a hill at 85 MPH and thus preventing truck versus car accidents. The systems may even be good at sensing the presence of a child riding a bicycle on the side of the road. But what about more unpredictable things, such as a motorcyclist who is weaving in and out of traffic or who is lane splitting? The ability to react to such driving is what can prevent a motorcycle accident. That reactive instinct may be only something a human truly has. It is difficult for anyone (or anything) to avoid an accident when a driver is doing dangerous things. Motorcycles are often difficult enough to spot even when they are obeying traffic laws. How do the systems evaluate the best path to take in trying to avoid a collision? Humans might make the split second decision to steer towards the empty car to their left rather than the motorcyclist to their right, but a machine might interpret the smaller mass of the motorcycle as being the better option to steer towards.
If driverless cars become a reality on U.S. roads, the issue of accident liability will arise soon enough. Before even approaching the liability issue, it would be wise to ascertain the effectiveness of the driverless systems, prior to the public placing their trust in them.
Updated: A few days ago, one of Google’s driverless cars was involved in an accident. Ironically, this apparently happened after the driver had switched to manual control.
Related posts:
