Collaborative Situation-Awareness in Vehicles by Means of Spatio-Temporal Information Fusion With Probabilistic Networks

Predictive driver assistance systems are a key issue in the visionary field of accident-free driving. Vehicles should be enabled to foresee critical driving conditions and therefore inform their drivers in a timely manner. In this context, direct communication among vehicles complements on-board sensor systems, enabling a new paradigm in driving assistance: collaborative situation-awareness. However, such predictive and cooperative systems typically have to deal with incomplete and unconfident information. In addition, spatial and temporal effects have to be taken into account. In this article we present a novel approach for collaborative situation-awareness in vehicles based on the probabilistic network paradigm. The approach enables a spatio-temporal analysis of remote locations ahead by using additional intermediate nodes, thereby utilizing both, on-board sensor data and previous observations from other vehicles. We evaluate our techniques by simulation of a distributed heavy rain detection scenario