Automated road passenger transport
For a long time the media have been talking of self-driving vehicles as a futuristic solution for all kinds of transport problems. However, in recent months such plans have assumed a more tangible form, making an automated transport network even more likely in the future. And there is a particular focus on public transport.
Recently the Swiss community of Sitten extended the test phase of the so-called SmartShuttle, a pilot project in Estonia is already running its self-driving buses in the country’s capital, and last June a Chinese train manufacturer presented a tram, which not only operates without a driver but without rails too. Germany has also been trialling automation in this field for some time. In the spring of 2018 four fully automated minibuses are due to start operating on the grounds of the Charité hospital in Berlin, while the rail company Deutsche Bahn plans to use driverless e-shuttles to carry its passengers from their own front doors to the nearest station by 2015 at the latest.
Using cameras and sensors, the automated buses are operating in a preliminary phase on designated routes and, according to Hans Sax, head of the Karlsruhe Institute for Technology, systems could also be installed to exchange details about the number of people on board and about weather or road conditions. However, Sax believes that certification for individual, self-driving vehicles is still a long way off.
However, an example of the sort of number of scenarios that are conceivable in the next few years is provided by Daimler’s “Future Bus”, which is operating a shuttle service in Amsterdam between Schipol airport and the suburb of Haarlem. Although this bus is self-driving, can recognise pedestrians and red lights, and automatically allows passengers to board and alight, it does not entirely dispense with personnel. There is still a driver in the cab to monitor the system and intervene in an emergency. This is an example of the third of five stages of automation to which technical systems in this field can be assigned. Whereas the first two stages, assisted and partially automated driving, for example in the form of cruise control and autopilot, can already be found on today’s roads, stage 5, a fully self-driving car without a steering wheel, is inconceivable at the present time.
Therefore the next stage, which is likely to be a major determining factor on transport networks over the next decade, will be Stage 4 of this system: fully automated, roadworthy vehicles in which drivers are present but do not need to concentrate on the traffic or intervene in the driving process. The current study by Roland Berger assumes that vehicles with this degree of automation will be a regular feature of the automobile industry by 2030, and anticipates some significant changes in mobility based on this assumption.
The most far-reaching effects of such developments will be in public transport structures, where a reduction in expenditure of up to 60% can be expected. This will be mainly a result of the need for fewer personnel.
However, this is accompanied not only by the potential expansion of the transport network as a result of newly acquired capacities, for example, the use of automated buses operating between city centres and surrounding rural areas, but also by concerns about the loss of a significant number of jobs. At the very latest, the continued development from stage 4 automation to stage 5 would instantly jeopardise the job security of millions of professional drivers.
Projections regarding the automation of commercial vehicles reveal that by 2030 some 50-70 per cent of commercial drivers in the USA and Europe will be surplus to requirements. A similar effect could potentially arise among professional drivers in the public transport sector. However, with suitable retraining, it is conceivable that at least some of the workforce could find employment in other areas of the transport system, such as in the maintenance and monitoring of automated operations.
How imminent such changes to the structure of public transport are depends, however, not only on the continued development of the technology used, but also on the accompanying costs incurred not only by vehicle manufacturers but also by the users of automated forms of transport.
The savings on personnel costs might not necessarily impact positively on public transport. According to Roland Berger, price changes could lead in particular to an increase in the fleets of self-driving taxis, and this could substantially reduce the appeal of public transport to potential passengers. Furthermore, more taxis and a concomitant decline in the current availability of car-sharing could increase the overall amount of traffic. One consequence of this would be longer journey times for buses operating public transport services. Moreover, as a result of lower demand, public transport users would not enjoy any significant cost reductions.
Whereas the expense involved in building up fleets of self-driving taxis would be confined to the cost of purchasing the vehicles, the automation of the public transport network would generally be accompanied by the expense of converting the infrastructure. This applies in particular to the rail network, including trains and metros. Although the construction of a driverless metro line only results in a three per cent increase in costs compared with the conventional version, the conversion of existing routes would add some 30% to the cost. This puts the savings in personnel costs into perspective, and in such a case the conversion to automated systems would not appear to be particularly worthwhile either for the public transport operators or for their users.
It is a similar situation with regard to the automation of private cars: whereas driver assistance systems such as proximity sensors or traffic jam assistance are already standard equipment in many new cars, at the present time automated systems are not commercially viable either for the manufacturers or for potential buyers. The simple reason for this is the cost of the required technology. For example, its installation in the current test vehicle being used by Bosch costs around half a million euros. At the present time consumers are willing to dispense with a highly automated system when confronted with such costs. Manufacturers also have to take into account the additional costs incurred by testing self-driving cars, whose safety can only be guaranteed by conducting several million kilometres of testing.
In terms of traffic safety the vision of automated vehicles is not on very firm ground. An international survey conducted by Deloitte, which also involved 1,750 German consumers, revealed that we are still not confident about allowing our vehicles to be driven by machines. For 72% of German participants the safety of self-driving systems is seen as inadequate and they expressed serious reservations about them. Only 20% of the mid-range age group would theoretically consider using an automated vehicle. The younger generation were slightly more enthusiastic. The willingness to accept self-driving vehicles among so-called millennials – i.e. those born between 1980 and 2000, was just under 30%, while for Generation Y, born between 1995 and 2010, it was just over 30%. It is a fact that, when considering the development of futuristic systems, safety-related as well as ethical aspects have to be given adequate consideration if a far-reaching change in transport is actually to take effect.
How would such a system react, for example, to unpredicted situations? Even though automated vehicles are likely to be the cause of fewer accidents than those due at present to human drivers, nevertheless they will not be entirely accident-free. The Massachusetts Institute of Technology has developed an online experiment by the name of Moral Machine, which clearly shows the potential dilemmas involving self-driving vehicles. When in doubt, the car has to make a life or death decision. For example, if a vehicle approaching a pedestrian crossing experiences brake failure, and avoiding the pedestrian is certain to cause the death of the occupants of the car, neither a human being nor a machine is in a position to deal with the moral implications of such a decision. At what point can or should a human driver intervene? And if an accident occurs, who bears the liability? In the middle of last year the Ethics Commission chaired by Udo Di Fabio produced twenty principles for automated driving systems, making Germany the first country in the world to create the relevant guidelines for the use of driverless vehicles. One of these principles states that cars should only take over the controls on the highway if this can be demonstrated to improve road safety.
The mathematical model currently being developed by the Israeli manufacturer of assistance systems, Mobileye, could be one approach for unequivocally guaranteeing this increased level of safety. It aims to prepare self-driving systems for potential accident scenarios by using a mathematical formula, and in this way to develop a standardised driving response to emergency situations. Even so, the question of how to deal with unexpected situations when driving remains unresolved.