Decision fusion of a multi-sensing embedded system for occupant safety measures

The need for an embedded system that can fuse incomplete, inconsistent, and imprecise decisions from several sensing systems is a crucial step in achieving an effective decision for occupant safety measures. This paper deals with the decision fusion strategies of a multi-sensing embedded system to achieve significant enhancement in the reliability of occupant safety through the fused decisions. Multi-sensing approaches to determine weight, vision, and crash sensing are developed for occupant detection, classification, position calculation, and crash detection. A rule-based decision fusion algorithm is then developed to fuse the multi-sensing decisions. The developed sensing systems are incorporated into an embedded device. To execute the embedded system, a system interface between the software and hardware is developed using Lab Window/CVI with the C programming language. The experimental results demonstrated that the real time operation of the embedded system validate the effectiveness of the decision fusion algorithm, characterize the safety measures and monitor the decision application. Several events were tested that prove the performance of the embedded system is robust towards occupant safety measures.