Every so often a technology comes along that makes a fundamental change to a given field. Like MP3s to the music industry. I am convinced that self-driving cars (also known as autonomous vehicles) is such a technology for transport. New technologies always bring both pros and cons. Do not believe anyone who tells you that self-driving cars are a nightmare scenario. And do not believe anyone who claims they are a panacea that will solve all our transport problems. They will bring a spectrum of impacts.
Self-driving technology is not yet sufficiently developed to allow fully autonomous vehicles on our roads today, but we are probably within five years of a small number being present, and within ten years of more widespread use. The next decade has the potential to see a revolutionary change in how we view cars, roads and transport in general. It is impossible to do the whole area justice in such a short space, but in this blog I very briefly summarise some of the potential changes that are coming:
Liability for crashes will move from driver to manufacturer.
If you are behind the wheel of a car that drives itself, who is responsible in a crash? Clearly it could not be the occupant if they were not in control of the vehicle at all, and was not required to be. So liability will have to lie with the manufacturer of the self-driving hardware and software.
We will no longer have to have a driving license.
A self-driving car will be capable of driving anyone, whether or not they can drive a car. In theory a child under the age of 17 could travel alone, as could a person registered blind. There will no longer be a need to take driving lessons, pass a driving test or hold a driving license.
We will no longer own our cars.
If the car contains all the hardware/software for driving the car, and the manufacturer is liable for crashes, then they will not want to hand ownership of the car over to someone who could compromise the technology. So it will be more like software where we use the car under the terms of a license, rather like a rental agreement. So you will hire the car in your driveway, not own it.
But why stop there? If a car can drive itself, why go to all the bother of having one sitting in your driveway taking up space? Why not just have one turn up when you need it and have it drive away again when you’re done? It may not be the same vehicle each day, but is that really a problem if you just want to get from A to B?
There will be different models of car ownership.
There will always be car lovers so there will continue to be a demand for manually controlled cars. However, in time the cost of getting a driving license for a manually controlled car will increase as their numbers decline. And insurance costs will sky-rocket since human drivers will be the most dangerous on the roads, compared to computers. So the top end of the hierarchy will be the rich who can still afford to drive manually controlled cars.
Below this are people who are happy to hire a car, but want it to be “their” car, so they can fill it with their stuff and have it in their driveway, or if parked off-site at least be the same vehicle each day. This will operate like a hire-purchase arrangement.
Below this, at lower cost, will be those who are happy to hire a car, but don’t mind if it’s a different vehicle each time they use it and don’t mind if it’s not parked outside their home. They may pay on a pay-as-you-go basis or via monthly subscription.
The bottom rung will be people who are also happy to car-share with strangers. They will pay even less in return for agreeing to perhaps not be taken directly to their destination and to accept the car picking up and dropping off other people on the way. This is not dissimilar to a black taxi today.
The distinction between public and private transport will blur and even disappear.
The bottom rung of people who hire cars and share with other people is not dissimilar to a bus, except that the route is bespoke and the vehicle is smaller. But there is no reason why self-driving buses could not also appear. These, too, may be owned by manufacturers. This raises the possibility of a more nuanced transport system with a descending hierarchy of large buses, small buses, large cars and small cars all serving progressively lower-demand journeys. It will be hard to tell where “public transport” ends and “private transport” begins – and will the distinction even matter?
Cars will have sorter lifespans.
If cars are hired out, manufacturers will want to make sure they get as much use as possible. This means that they will incur mileage much more rapidly than cars which are parked outside homes all day. So the average lifespan of a car may reduce to less than five years, meaning that the majority of cars on the road may be quite modern. This has the advantage that new technologies will be disseminated much more rapidly. There will also be a need to maintain all these cars, and this could be a new industry, or a development of the existing car maintenance industry. Car usage patterns suggest that it would make sense for a lot of car maintenance work to take place during the night when cars are idle.
Traffic levels could rise significantly.
At the present, the number of drivers limits the number of cars on the road. Logically, there cannot be more cars on the road than there are licensed drivers. This limit disappears with cars that can drive themselves. With car companies sending unoccupied cars out onto the roads, traffic levels are limited only by the number of vehicles that exist, which could be much higher.
Car parks will move from where people work/shop to the city periphery.
Currently people drive to work and shops and obviously seek to park close by. This creates demand for large car parks in city centre locations and at out-of-town shopping centres. But with self-driving cars there is no need to park close to shops or workplaces. Instead, there may be a need for “drop off points” where cars can let out passengers, before driving away to get other passengers. However, demand will be higher at some times (eg 8-9am) than other times (eg 3am) and so at off-peak times there will be a need for somewhere to store thousands of idle cars. Economics demands that these be in areas where land is cheap, namely industrial estates or the city periphery. So we will see a loss of demand for city centre car parks and an increase in demand for large-scale out of town parking areas. This reduction in parking could make both housing estates and city centre streets much more attractive by removing parked vehicles from the streets and freeing up land currently occupied by car parks, and allowing increases in density of development.
Almost perfect compliance with the law will be possible.
Self-driving cars will drive exactly as programmed and will not experience human emotions such as frustration. As such, almost perfect compliance with things like speed limits and traffic signals becomes achievable. Currently legislation is limited by the practicalities of enforcement – where even if a law is passed, it cannot be enforced. But imagine a scenario where you could establish a 17mph speed limit outside a particular school, and have immediate compliance by almost all vehicles. Or mark a particular route as “residents only” or “not suitable for HGVs” and have the same immediate compliance? Or even create laws that are impossible to pass today, for example one that says “traffic going from Lisburn to Bangor must use the A55 Outer Ring”. It would allow governments to very closely control how roads are used for the first time.
Having a small number of manufacturers supplying vehicles would allow governments to much more closely regulate the use of vehicles. For example, governments could use a taxation system, based on miles travelled, time travelled and location travelled to create a complex marketplace for transport with the aim of influencing where people drive to and when. So a higher taxation on inner city streets would seek to discourage traffic build-up in sensitive locations, while lower taxation to areas of social deprivation could be used to encourage investment.
HGVs will also be increasingly self-driving. This will make freight transport cheaper since there will no longer be a driver’s wage to pay, and HGVs will no longer need to stop for rest breaks or have tachographs. Moving goods about will be easier and cheaper. Like cars, there will be a hierarchy of HGVs, smaller lorries, vans and courier-type vehicles. Perhaps companies such as Amazon could run a fleet of self-driving delivery vehicles.
Significant reduction in road deaths.
Nobody is suggesting that self-driving cars will never make a mistake or never crash. But they do have the potential to be significantly safer than human drivers. A plausible figure that is sometimes quoted is “ten times safer”. This means that for every ten people killed on the roads driving manually-controlled cars, only one person would be killed if they had been using self-driving cars. Getting road fatalities in Northern Ireland down to seven or eight per year would seem to be within reach with self-driving technology.
Much safer for pedestrians and cyclists.
Self-driving cars have already demonstrated strong awareness of what other road users are doing. A Google test car once did an emergency stop because it thought that a cyclist on a footway, who did a slight wobble, was about to fall onto the road. This may have been an over-reaction by the test car, but it illustrates the point that a computer can be thinking about the trajectory and possible actions of dozens of other road users simultaneously. Cars could potentially spot a child about to run into the road and take action before they have even reached the road. It is possible that at some future date a pedestrian could simply step off the pavement into live traffic in almost perfect safety. This is something that could certainly be abused by pranksters, but also allows for the possibility of streets becoming much more equitable between different types of road user.
Roads would have much higher capacity.
Self-driving cars currently leave large gaps between themselves and human-controlled cars, but once the majority of cars are autonomous, they could safely drive much closer together. Imagine the M1 into Belfast filled with cars travelling 2 or 3 metres apart at 50mph. The capacity of the road network would rise significantly without having to build new roads or widen existing ones.
Hacking is a threat.
Self-driving cars could potentially be “hacked” by terrorists. A “hacked” HGV with its safety features over-ridden could be driven through a pedestrian area much as occurred in France in 2016. But in this case, there would be no driver to incapacitate, making the incident much more difficult to bring to an end. Making autonomous vehicles safe from hacking will become a significant issue.
Lifestyle changes and less of a concern about journey times.
Currently a commuter obviously has to focus on driving their car. But if cars are self-driving, then the occupant can do whatever they wish with the time. This could include eating, sleeping or working. So what was once “wasted” time commuting could now become productive or leisure time. This means that commuters may become less concerned about their journey times and hence may be less concerned about congestion than at present.
It would also make longer commutes more tolerable. Imagine if you lived in Belfast and worked in Strabane. You could get up at 7am and get straight into a car with a box of food, go back to sleep for the first hour and then eat breakfast and catch up with the news or social media during the second hour before turning up at work at 9am. They would arrive in work less tired and stressed.
Some professions will disappear.
Some common professions will ultimately no longer be needed, or needed in fewer numbers. These include bus drivers, taxi drivers, traffic wardens, traffic police, private car mechanics, tyre shops for the general public, driving instructors, driving test officers and lorry drivers. Professions have come and gone throughout history but of course cause problems for the individuals affected.
The final 5% is the technological hurdle.
The technology exists today to introduce self-driving cars that can drive in all normal driving conditions. The delay, and the majority of the work, comes from dealing with the exceptions. What, for example, does a self-driving car do in these circumstances?:
- A police officer steps into the road and indicates “stop” with their hand.
- A manhole collapses in the middle of the road ahead.
- A fallen tree is blocking the road.
- Someone alters a road sign to show the wrong speed limit.
- A pedestrian sits down in the middle of the road.
- The car suffers a mechanical failure on a motorway.
- The road markings are covered with snow.
The unexpected is by definition unexpected, so the cars will somehow have to learn and adapt and this raises the potential for errors. Self-driving cars will therefore make mistakes and we need to be clear about this. It is this “final 5%” of situations that is going to take 95% of the work for those developing the technologies, but the problem will eventually be solved and self-driving cars will arrive on our streets. I believe it is inevitable, and it will be up to society to meet the opportunities and problems highlighted in this blog and adapt.
Wesley Johnston, 7 Jan 2017, http://www.wesleyjohnston.com/roads