During this Festival, a small autonomous bus is driving some hundred meters on a dedicated lane between an exhibition and the entrance to the Festival neighbourhood. Residents and visitors are queuing up in long lines and seem to be highly interested to sit in such a vehicle. This new technology was also heavily debated during the EcoMobility World Congress (2-4 October 2017) in Kaohsiung, see the “Kaohsiung Strategies”, an outcome document which takes up new challenges to urban mobility, among these autonomous driving. The “Shared Mobility Principles for Livable Cities” were also launched in Kaohsiung.
Time to reflect about the pros and cons of this technology which is pushed into our cities. Today, many driving assistance systems are already built into modern cars. A whole range of such systems actively steers, accelerates or brakes, and provides collision warning, pedestrian recognition and automated parking. These new technologies aim to increase convenience for drivers as well as road safety.
Less traffic and reduced fossil fuel consumption: arguments in favour of autonomous driving?
But what are the chances and risks of autonomous driving? To carefully assess them, we need to look at both sides of the discussion. Here are the arguments we usually hear from the advocates of a positive view of autonomous driving:
Autonomous cars can help to reclaim urban spaces. Successfully reducing cars on the road by making use of shared autonomous vehicles would allow cities to reorganise streets and to reclaim urban spaces currently used for parking. Thus, new opportunities would arise for converting street space into living space, while avoiding the huge costs of providing parking lots.
Autonomous vehicles allow for new sharing models. Carsharing is increasingly common. Technological tools such as smartphones allow for innovative platforms that attract new business models using shared vehicles. With the advent of autonomous vehicles, carsharing has the potential to become even more convenient than owning a private car at just a fraction of the cost. The motto is: Use it, don’t own it.
Autonomous cars will lead to fewer cars on the street. Urban transport modelling shows that about 10 per cent of today’s vehicle fleet would be sufficient to accommodate the motorised transport needs of urban populations if turned into a carsharing model. New types of vehicles and autonomous fleets, which are more easily accessible compared to conventional carsharing systems, could accelerate the trend towards sharing.
Autonomous vehicles are compatible with other modes of transport. High capacity collective transport systems such as metros, light rail and bus rapid transit (BRT) can maintain their relevance and would not be endangered. If public transport companies are sufficiently innovative and fast, they can open new markets and provide new flexible services integrating autonomous vehicles into existing mass public transport options.
Autonomous cars can help reduce fossil fuel consumption. This argument assumes that a new generation of autonomous vehicles will be electric. Additionally, cars will be able to organise their routes according to battery status and will drive automatically to recharging stations. Provided the electricity comes from renewable energy sources, CO2 emissions and air pollution would be reduced.
Autonomous cars improve road safety. Currently, more than 90 per cent of road accidents are due to human error and risky driving behaviour (such as driving under the influence of drugs or alcohol, or speeding). Autonomous driving could reduce that risk.
In short, these arguments describe autonomous driving as a brilliant solution to many urban transportation challenges.
Increased transport volumes and separated urban roads: the risks of autonomous driving
Autonomous driving technologies include risks and not all arguments in their favour are correct. Risks and negative impacts are rather obvious:
Autonomous cars may drastically increase transport volumes. Fully autonomous cars could function as electronically connected “driverless taxis”, easy to call and cheap in their use. This may actually lead to an increase in mileage travelled instead of a decrease. In addition, the so-called rebound effect (= efficiency gains are compensated through higher demand/ use/ consumption in numbers or volumes) bears the risk of taking away the advantages of more efficient transport operation with electronically connected cars.
The following are only some of the reasons that could lead to an increase in transport volumes:
- Modal shift towards autonomous cars: As the “driverless taxi” offers door-to-door transport without today’s pain of finding a parking spot, such ultimate convenience may lead to a shift away from sustainable modes of public transport, walking and cycling towards driving.
- New user groups: Using an autonomous car will be as easy as using an elevator today. No driving license will be required, you will only have to give the destination and identify your payment method. This could be more inclusive and offer more independence for instance for elderly people, thereby creating a positive benefit. However, it may result in substantially more trips in total by users and by autonomous cars themselves, for instance if families send their kids to school or their cars to the bakery.
- Travel time is no longer “wasted time”: Today’s transport modelling builds on an average constant travel time which people are ready to invest. With no need to concentrate on road traffic, the user of an autonomous vehicle can enjoy entertainment, rest or work. Such a new disconnect between travel distance and travel time can easily lead to acceptance of longer distances travelled by car.
- Convenience has impacts: By reducing the inconvenience of travel time, one can ultimately cause further urban sprawl as it becomes more appealing for people to move out of (inner) cities, e.g. to rural areas and remote places. Consequently, the kilometres covered to reach work, education or leisure go up and city boundaries further blur. Hence the principles of compact cities and regional planning goals are heavily endangered.
Autonomous driving may lead to more separated urban roads. Autonomous cars are programmed to drive defensively, stopping or slowing down for all “obstacles” like crossing pedestrians. While this is helpful to make urban streets safer and more liveable, opposition to such “traffic calming” might come from interests for high speeds on major urban roads.
Whereas some technology developers are calling for fencing off such major roads to clearly separate sidewalks from driving space and prevent pedestrian from crossing, urbanists fear such an approach and are looking look for more comprehensive solutions. Only solutions that aim to keep driving speeds low and use instruments such as financial incentives to limit cars on the street can avoid further physical and social separation of cities. Fenced urban streets cannot be a goal of sustainable urban development.
Autonomous vehicles pose risks to privacy and cyber security. The use of autonomous vehicles will generate a huge amount of data. Who will own this big data and what are the consequences of the continual monitoring and documentation of each and every vehicle and its moves? And, as cyber-attacks become more frequent, can cyber security be ensured? How can the misuse of autonomous vehicles as weapons be avoided?
Autonomous e-vehicles will cause unforeseen problems to public income. As long as tax systems are not reformed, e-vehicles would not pay fuel taxes and thus would not contribute to covering costs anyway. But even more relevant for the local level: autonomous vehicles hardly park anymore (they rather circulate on the streets!) and thus a relevant income from parking fees will be missing.
Many unanswered questions remain
The transition towards autonomous driving is complex, and we are already in the middle of it. An increasing number of new cars are prepared for assisted or partially automated operation, which leads to key operational questions:
How much of the current traffic regulations will remain intact? Will the vehicles be programmed to drive defensively and follow all existing rules including speed limits, or will they be able to speed up in automated mode? The story of a Tesla driver who crossed the US in automated mode in record time, apparently not having kept given speed limits, gives reason for concern.
How will the combination of automated and human-steered cars work if they interact on the same roads? The gain in efficiency, which is a strong argument in favour of automated vehicles, can only be fully exploited with electronically connected, computer-controlled cars. Human drivers are a “disruptive factor”. There may come a time when a ban of non-automated cars for certain areas or highways will come into discussion. Is that something we want?