An analysis of synergy effects in an omnidirectional modular perception system

Abstract

The increasing complexity of Advanced Driver Assistance Systems (ADAS) requires new design strategies in order to provide the necessary reliability for safety-related applications. In this paper, the connection of several perception modules to a modular ADAS is proposed. The benefits of such an approach are analyzed on the example of a vision and radar based vehicle & lane detection system. In particular, the effect of algorithmic synergies between the modules is investigated. The performance of the modular perception system is demonstrated using real-world data from the concept vehicle Carai. The results show that the performance and reliability of this approach surpasses the properties of the single modules.