Driverless Cars

The driverless cars developed by Stanford University are an innovation to behold:

It is easy to understand how driverless cars would be safer.  Maintaining a constant speed on the motorways will reduce chaotic braking, and the small variations in spaces and speed that create phantom traffic jams will be eliminated.  And with a little bit of linking technology between a group of driverless cars, hazards in one place can be communicated to the other cars on the road much quicker than human drivers with their relatively poor reaction times.

Koushik Dutta runs through some of the implications of a driverless car:

The operating percent of a car will go from 4% to that 96%. But back to my leading statement: there are unintended consequences. Parked cars will be a relic from the past. What happens to car insurance prices if a driver is no longer part of the equation? And if cars are receiving 20 times more actual use, that would imply that there would be 20 times less cars sold. This is the kind of disruptive change that can reshape the automotive industry. The recent GM/Chrysler bailout may have been for naught. … Of course, car companies realize this. And I can guarantee you, they will lobby against driverless cars.

Despite the clear benefits of such driverless technology, one can also see how winning changes in legislation to operate driverless cars may lag far behind the technology. Aside from the active lobbying against such schemes from car companies (and haulage unions, taxi drivers, &ct) I imagine people and legislators would be slightly squeamish about letting automated cars out onto the road.  Even though we know that auto-pilots do most of the commercial airline flying, there is something reassuring about the fact that the pilot is on board, sharing the ‘danger’ of flying with you.  Presumably, a passenger in a driverless car would be able to take control of the vehicle if they needed to… but the real benefit of such a car is precisely that it can drive itself home (or, come and pick you up), making a portion of the journey with no-one in the vehicle.

There is also the problem of mixing human driven cars, with driverless cars.  The safety benefit of the new technology is surely at its greatest when everyone is using the driverless technology.  All vehicles can travel at a constant speed and there would not be any crashes.  But in order to introduce such technology, and to get it widely adopted, you have to go through an intermediate stage where early adopters have to share the roads with the human muggles who still insist on actually driving their cars.  Perhaps legislators will demand that the driverless vehicles are specially painted, or have flashing lights on them, to warn other drivers that their is something on the road that will not behave in the way you might expect, much like ‘Long Vehicle’ and ‘Wide Load’ livery on haulage vehicles.

A related problem is that when there is no actual person in the vehicles we share the road with, the moral duty we feel we owe to other drivers to stay safe will dissipate.  Driverless vehicles will not be given right of way, and human drivers will cut in front of driverless cars more frequently.   Young joyriders on bikes or in cars could start ‘teasing’ the driverless vehicles, deliberately driving erratically to test the avoidance capabilities of the software.

There is also a civil liberties concern, in that driverless cars will presumably log every journey they make somewhere, for diagnostic and ‘learning’ purposes, but this information could be exploited by the state, companies or anyone else who wants to invade your privacy.  Governments or commercial interests could programme cars to refuse to take you to certain locations, or to drive you via advertising hoardings.  This would be undesirable… but appropriate technological checks could easily guard against such abuse.

The way to introduce such technology is in a closed system, where the entire road infrastructure can be controlled. The DLR operates without drivers, and a new pod system has been introduced at Heathrow Airport, where driverless pods operate on dedicated lanes. Perhaps Heathrow or another airport, one out of the city centre and with a spur road serving it, could invest in dedicated driverless lanes, plus detailed road mapping, and some sort of API for their traffic lights? This would allow driverless cars to operate efficiently to-and-from the airport, and provide a ‘proof of concept’ to legislators and regulators.

Finally, there will be car enthusiasts who insist that driving a car is one of the joys of life. Why surrender it to a machine? Well, yes, but even though horse riding was made obsolete as a system of mass transit when engines (steam, internal combustion) were invented, enthusiasts can still do it for fun. But for those who only drive out of necessity, driverless cars offer a tantalising glimpse of a congestion free future.

4 thoughts on “Driverless Cars

  1. How amazing. If there’s a passenger in the car, (esp if it’s got one of those Baby on Board signs), then surely there’s still a moral duty to stay safe (not to mention a self-preservatory one). I’d imagine the problem would be more like what cycle helmets did for cycler safety – ie if people know there’s a protective mechanism in force then they might abdicate responsibility to the mechanism and take more risks themselves.

    Very amazing though. It seems to me, the more technology advances, the more imperative there is to find a technology to eliminate self-seeking, corrupt or oppressive government/commercial practices.

  2. I wonder if that would be possible. If policing decisions, say, were automated, there wouldn’t be police corruption. If taxation policy (and spending decisions) were automated according to a fairness algorithm, there wouldn’t be disproportionate assaults upon the poor and vulnerable. And if law-making were automated, we could maximise the distribution of fairness equally across society. Could it in theory be possible to remove the incentives from politicians for bowing to business- and self-interests? Yes, that’s it! I want a driverless government.

  3. What if there is a technical failure of the Google system while on a busy motorway? Many well known motor manufacturers have been know to withdraw various models due technical defects.

    1. One assumes that there would be procedures in place to mitigate any failure. First, if the passengers are not driving, then the seating orientation of a car does not need to be waht it is now. Cars could be designed to give passengers greater comfort. Second, a passenger could take over the driving if the system failed. Completely passengerless cars (i.e. those where no human could take over in the event of failure) would be sent along different channels or ‘travel windows’ to those with people in them.

      But I guess the wider to point to make is that because technical failures would be rarer than driver error, the aggregate number of accidents and fatalities would decrease. The fact that aeroplanes have occasional catastrophic failures, leading to loss of life, is not an argument against the use of those systems in air travel.

Leave a Reply