Predictive risk maps

This paper addresses the problem of risk assessment in dynamic traffic environments for future behavior evaluation and planning. Risk assessment has to be driven from behavioral needs of the acting scene entity to evaluate the best possible future behavior. We present a general approach for a temporally continuous future risk estimation. Using this risk estimation, we introduce predictive risk maps based on the variation of the acting entities´ possible dynamics. The predictive risk maps indicate how critical a certain behavior will be in the future and we use them for future behavior evaluation and planning. Explicitly introducing risk indicators for the collision case of moving entities we show the generality of our approach by applying it to several different traffic scene types.