Research on Virtual-Physical Integrated Testing Methodology for USV Collision Avoidance Safety

This study addresses the testing requirements for evaluating whether the collision avoidance success rate of an unmanned surface vessel (USV) meets the specified threshold during autonomous collision avoidance trials. First, the USV’s onboard intelligent navigation control system (including its intelligent navigation software)—a critical component influencing collision safety—is selected as the test subject. Subsequently, a testing framework is proposed to progressively transition from virtual simulation to physical testing. Comprehensive test cases, generated through multidimensional conditional space search, are executed on a lab hardware-in-the-loop (HIL) simulation platform, after which key test cases are screened and validated in field-based physical trials. Finally, a testing toolchain compatible with virtual-physical integrated experiments is designed to ensure effective implementation. The research resolves challenges such as the inability to conduct extensive physical tests due to time and cost constraints, high operational risks in real-world trials, and the difficulty in defining confidence levels for virtual simulations caused by limitations in digital model accuracy.