There are nine completely driverless train systems/lines operating in Europe, eight in Asia and six elsewhere. There are a further nineteen in Europe with a “standby driver” or, like London’s Docklands Light Railway, with a “Passenger Service Agent” present on the train, just in case something goes wrong.
So Google’s claim that its seven driverless test cars have driven 1,000 miles on roads without human intervention and more than 140,000 miles with only occasional human control sounds plausible. The company is reported by the New York Times as saying one car drove itself down Lombard Street, one of the steepest and curviest streets in San Francisco.
According to the paper, Google’s engineers say “robot drivers” are better because they:
React faster than humans, have 360-degree perception and do not get distracted, sleepy or intoxicated……They speak in terms of lives saved and injuries avoided — more than 37,000 people died in car accidents in the United States in 2008. The engineers say the technology could double the capacity of roads by allowing cars to drive more safely while closer together.
Although they are some years away yet, the claimed potential benefits of this new technology are enormous. If proven, it should allow travellers to do other things while driving, making time spent travelling much more productive. On roads where conventional vehicles have been superseded, road capacity should at least double, although according to some observers an eight-fold increase can easily be achieved. Speeds should increase while simultaneously reducing road accidents — one of the largest negative externalities associated with roads – through keeping drunk drivers away from the wheel and minimising simple driver error. If accidents are less likely, vehicles can be made lighter and therefore use less fuel.
If it can be implemented without the need for a “standby driver”, there is scope to lower taxi and freight costs substantially. In the latter case this should help make smaller trucks viable, reducing the need for very large trucks within urban areas. However the natural extension of eliminating the need for drivers is to remove the requirement to own cars altogether. If all the functionality of a private car is still possible – like on-demand availability, privacy, point-to-point travel – then the warrant for owning a dedicated vehicle is greatly reduced.
Huge benefits would follow if sharing could be made to work because rather than being parked for 98% of the day, vehicles could be out earning their keep 24/7. The size of the city’s car fleet would be greatly reduced and the cars themselves could be much smaller and lighter – for example, a majority could be single seaters to reflect demand patterns. In time, it’s likely the cost of travel attributable to vehicle ownership and fuel costs would fall significantly as economies of scale were achieved. People on lower incomes or unable to drive would get a big improvement in mobility. Travellers would ‘pay per kilometre’, making them more sensitive to travel costs.
For all the promise of driverless cars, there are also downsides. There would be a long period when the fleet wouldn’t be completely driverless, potentially giving rise to conflicts with conventional vehicles. Some people simply like “driving”, some might still resist sharing vehicles and some are bound to be suspicious about entrusting their personal welfare to new technology with potentially life-threatening implications in the event of a malfunction. Travellers would rightly want to feel they were fully protected from any risk of hackers and terrorists compromising software. The network effects of driverless cars are probably not well understood – it’s possible the benefits might not be as large as anticipated.
Tyler Cowen thinks the key obstacles to take-up of driverless cars in the US won’t be technical but rather social and political. He points out they are illegal at present in all States because the law assumes the existence of a driver who is responsible for the conduct of the vehicle (Google always has a “standby driver” at the wheel). He also thinks politicians would find it impossible to avoid clamping down on driverless cars in the early years in the event there were a serious accident (which I think would be inevitable):
There could be demands to shut down the cars until just about every problem is solved. The lives saved by the cars would not be as visible as the lives lost, and therefore the law might thwart or delay what could be a very beneficial innovation.
The historical experience is that lower travel costs induce people to live further from major activities, giving rise to greater sprawl. However if access to cars is universal; if the size of the fleet is smaller; if cars are vastly more fuel-efficient, environmentally friendly and cheaper to operate; then the grounds for opposing sprawl are significantly reduced. There’s an analogy here with the way new train lines in the late nineteenth century enabled families to begin the long era of suburban settlement (more than 90% of Melburnians now live in the suburbs i.e. >5km from the CBD). Nevertheless, policy-makers who have other reasons to discourage sprawl could choose to increase the price of car-based travel, for example by pricing road space.
I expect widespread adoption of driverless cars would have major implications for public transport, although I think this is a murky area at present. Presumably suburban feeder buses as we know them now would be a thing of the past, replaced entirely by shared, driverless cars. Yet even with a doubling or better of road capacity, there would still be road congestion in key areas like the CBD, requiring rail-based public transport to enable high density locations to continue to function. Although public transport only accounts for circa 11-13% of all trips in Melbourne, a lot of the capacity gain from driverless cars would be lost if a large proportion of these travellers transferred to cars.
While it certainly won’t be easy, I’m not as pessimistic about institutional resistance as Tyler Cowen. Technologies that bring big net benefits are inexorable. I think driverless cars are coming — probably initially in a cut-down form — sooner than is generally appreciated (that’s a prediction!). I suspect it will be easier for governments to accept and in turn “sell” the idea to the electorate than congestion pricing, which is really the only other plausible way of whipping our big cities into shape. I’m not as sanguine about progress toward full-blown driverless cars as some others because I suspect the technical issues are far harder than Google is letting on (Lombard St is not actually a difficult test).
Even so, urban and transport planning agencies should be giving driverless cars a central and prominent position in their current thinking – implementation falls well within their customary 20-30 year planning time frames (what a fillip it would be for Victoria if the law were amended and Google were invited to trial its driverless cars on our roads!).
Update July 11 2011: will humans adapt to the driverless car?