The matter of how automated vehicles should behave in everyday road traffic has ethical implications. For many years, numerous aspects have already been at the focus of public discussion, although the situational dilemmas often mentioned in this connection are of little help: As thought experiments, they merely serve to create awareness for the role of ethical issues in the development of automated vehicles. Rather, developers are required to resolve specific ethical issues in their daily work. Scientists at the Technical University of Braunschweig are addressing all these matters in the funded project “Value-based Decision Making” under the leadership of Prof. Dr. Markus Maurer from the Institute of Control Engineering.
Automated vehicles will be safer than those driven by humans – this is a promise often made when it comes to integrating automated vehicles into public road traffic. But what are the challenges involved, and to what extent is this even feasible in view of the technological and economic considerations? The urgently needed discussions already begin with the question as to what the protagonists affected by the technology understand by the term “safety.”
In public communication regarding automated vehicles, a so-called “vision zero” expectation is often evoked. This would suggest that the technology is capable of eliminating all risks and enabling accident-free traffic. Engineers, on the other hand, define safety as the “absence of unacceptable risk.” Automated vehicles will always pose a certain residual risk that can be reduced, but not entirely eliminated. The nature of this residual risk cannot be determined by the engineers alone. Rather, what is needed is a well-balanced debate regarding the opportunities and risks of this technology that takes into account the concerns and interests of all parties involved. One goal of the funding project is to conduct precisely this discussion.
The residual risks at hand are brought about by a number of uncertainties that arise not specifically in connection with the technical systems, but the overall transportation system. Theoretically, anything can happen at any time: Road users do not adhere to the rules, people enter the roadway from concealed areas, and cyclists suddenly fall off their bicycles. Furthermore, the traffic system is constantly changing – for example with the recent addition of e-scooters as road users.
It is never possible to foresee or test all possible situations that an automated vehicle must be able to cope with in traffic – not even following the most conscientious development. In their daily work, the specialists must therefore make assumptions about situations that the driving system could possibly encounter in real traffic. These assumptions have a direct influence on the residual risk posed by the system and should therefore also be discussed together with the stakeholders concerned.
At the Institute of Control Engineering, approaches are being investigated as part of the funded project “Value-based behavioral decision-making” that describe the risk posed by automated vehicles in readily understandable terms. For this purpose, the scientists are using established systems engineering methods and semantic networks that make existing expert knowledge available for a technical system. The focus here is on clearly describing the assumptions made and the resulting risks in road traffic. In this way, the researchers are making an important contribution to transparent communication regarding the safety of automated vehicles.