Model-Based Design Approach for Validation of Vehicle Sensor Fusion Algorithm

Abstract

In recent years, the development of connected and autonomous vehicle (CAV) systems has accelerated dramatically as these systems can provide considerable benefits for vehicle performance and passenger safety. However, CAV systems can prove hazardous if not properly tested before deployment. A need for testing methods that are economical, time-effective, and low-risk has created the concept of model-based design, whereby systems are modeled and tested in various simulation environments before being fully deployed to the real system. This paper describes a model-based testing approach that has been developed to verify and validate a sensor fusion algorithm for CAV systems to enhance autonomous controls of a 2019 Chevrolet Blazer as part of the EcoCAR Mobility Challenge (EMC). Teams of undergraduate and graduate students with automotive interests laid out testing procedures to assess model fidelity and to identify and resolve issues with the algorithm before deployment to a student-adapted prototype vehicle. Students learned and applied complex engineering concepts rapidly to develop a sensor fusion and tracking algorithm.