A Test Method for Self-driving Vehicle Based on Mixed Reality

Abstract

This paper proposes a mixed reality-based self-driving vehicle testing method, aiming at self-driving vehicle safety testing in a complex environment. This method firstly maps the virtual scene in the simulation environment to the real test environment. Secondly, the performance indexes such as intelligent perception and the self-driving test vehicle’s behavior decision are quickly collected through sensor and data fusion. A closed-loop test system is then constructed based on the mixed reality test scene, control center, and test vehicle. Finally, through the case verification, it shows that this method can meet self-driving vehicles’ test requirements in various scenarios.