无人车硬件在环虚拟测试系统设计与验证

In order to solve the many problems faced in field road testing when verifying the intelligent behavior of unmanned vehicles, such as time-consuming, poor economy, high road environment requirements, and susceptibility to weather conditions. An unmanned vehicle hardware-in-the-loop virtual test system was designed to test the intelligent behavior of unmanned vehicles through virtual reality. The system included virtual traffic environment, road simulation, vehicle dynamics, and the perception of simulated vehicle sensors, and could meet the test needs of different traffic conditions, different models, and different sensors when testing unmanned vehicles. In the virtual traffic environment part, an automatic trigger mechanism for dynamic scenarios was proposed, which can be applied to the triggering of all dynamic moving targets in the virtual scene. In the vehicle dynamics part, two vehicle models were designed through the interface mapping scheme in the real-time system, The problems of low degree of freedom of the vehicle model and poor accuracy of the vehicle model in the scene software could be solved. In the simulation sensor perception part, two excitation signal schemes were proposed, on the one hand, and the deficiencies caused by the physical factors of the sensor target simulation system could be solved, on the other hand, it also better adapts to vehicle algorithm testing in complex scenarios. A virtual test scenario which designed by the system was used to verify the automatic trigger mechanism of the dynamic scenario in the scenario. The vehicle dynamics model of a real vehicle was tested, and the vehicle controller of the real vehicle was tested in the virtual test scenario. The results show that the automatic trigger mechanism can trigger the virtual target vehicle movement on time. The adaptive cruise capability of the vehicle controller of the actual vehicle to be tested is very good, and the following distance is stable at about 40 m in the low-speed scenario. In addition, the system has a certain degree of mechanical delay. Finally, the application of different excitation signal modes of the system shows that the black box simulation method has certain deficiencies in slope and curve scenes. The system can complete the virtual test of unmanned vehicle hardware in the loop.