车辆稳定性控制的虚拟实验

Vehicle Stability Control Approach Based on Virtual Reality

基于虚拟驾驶实验平台开发了车辆稳定性控制器。构建了由实车、虚拟场景、车辆动力学模型等组成的虚拟驾驶实验平台。用神经网络估计模型对车辆动力学系统进行了辨识,通过调整神经网络对象模型设计了神经自校正控制器,该控制器通过产生横摆力矩迫使车辆横摆角速度跟踪理想响应。在虚拟驾驶实验平台上研究了施加控制后车辆的响应及其与驾驶员的交互,对车辆在紧急避障操纵时的响应进行了虚拟仿真实验,验证了车辆稳定性控制器的有效性。

A neural network controller for vehicle stability control was developed based on the virtual reality platform. The virtual reality platform was constructed with real vehicle, virtual ground, vehicle dynamics model etc. The neural network approximate model was used to identify vehicle dynamics system and the self-tuning controller was obtained by rearrangement of the neural network object model. The controller was designed to generate additional yaw moment and make the yaw rate tracking the desired trajectory. Vehicle response and its interactions with drivers were investigated and virtual simulations of a vehicle under human avoidance maneuver were conducted on the virtual reality platform. The simulated results showed that the vehicle stability controller could improve vehicle handling and stability significantly.