ADAS Value Optimization for Rear Park Assist: Improvement and Assessment of Sensor Fusion Strategy

The project for this thesis consists in an ADAS Value Optimization activity conducted during an internship in Maserati S.p.A. with the objective of removing the ultrasonic sensors used for the Rear Park Assist (RPA) ADAS feature, obtaining the same functionality and performance in the detection and signaling of obstacles behind the car through a new system based on a sensor fusion strategy between Rear View Camera (RVC) and Blind Spot Radars (BSD). To achieve this goal, a study of the current RPA feature has been conducted, and starting from a previous implementation of the sensor fusion strategy for the new system, multiple updates and improvements have been implemented in order to achieve the functionality and performance required. Both hardware and software components of the system were updated and redesigned in the MATLAB/Simulink environment, and the final system obtained was tested through a standard validation procedure in a virtual simulation environment, obtaining encouraging results compatible with the RPA requirements and demonstrating the technical and economic feasibility of the developed RPA system based on a sensor fusion strategy between RVC and BSD which, after additional tests on the actual vehicle, could go into production.