Procedure and criteria for the safety of trials on public roads with (partially) self-driving vehicles : background and approach of the SWOV road safety advice. Vehicles are increasingly fitted with systems that take over (parts of) the driving task. In the most far-reaching form this will eventually lead to the fully self-driving vehicle. The human role will shift from driver to ‘supervisor’, and eventually to passenger. These systems are assumed to reduce the risk of human error and to therefore increase safety. At the same time, the human factor will still influence the systems. After all, in the role of supervisor, human intervention is still necessary on a system’s request or in the case of system failure. Furthermore, it is still unclear how other road users will react to the new systems. To stimulate innovations concerning self-driving vehicles, the Netherlands facilitates the testing of self-driving vehicles on public roads. As road safety is the main prerequisite, the Ministry of Infrastructure and the Environment asked SWOV for advice on how to carry out trials with self-driving vehicles in the safest possible way. The Ministry of Infrastructure and the Environment has drawn up a Procedure for testing Self-driving vehicles on public roads in the Netherlands; this procedure is used as a guideline when permission is asked for field trials on public roads. The procedure consists of three closely interrelated components: vehicle, road and human (behaviour). When assessing a field trial, RDW, Vehicle Technology and Information Centre, is responsible for the vehicle component, the road authority or CROW Taskforce Dutch Roads is responsible for the road component, and SWOV is responsible for the (human) behaviour component. The test procedure has been designed in such a way that improvements can be made based on the experience gained in each trial. The test procedure is therefore continuously under development. To give direction to the safety advice in the above procedure, the present SWOV report explores experiences with (partially) automated vehicles and with test procedures in other countries. The central question in this project is: How can we ascertain that tests with (partially) self-driving vehicles on public roads are carried out in the safest possible way?
Abstract