Abstract
Despite the significant results obtained during the past 20 years in reducing the number of killed or injured people in accidents, a new initiative (partly pushed by governmental authorities) is now undertaken to consequently reduce the total number of accidents. Thereby the most extreme vision is focusing on zero accident scenario, say a traffic situation not allowing any accidents to occur. It is believed that so-called driver assistance systems can significantly contribute to reach this goal in the future. The paper will focus possibilities and limitations. It is shown that these systems do require an adapting and self-explaining human-machine to allow the full deployment of their capabilities.