线控汽车驾驶员特性辨识算法的研究

A Study on the Identification Algorithm of Driver Characteristics for X-by-wire Vehicles

线控汽车相对于传统汽车具有较多的控制自由度,控制算法设计可将人的因素考虑在车辆集成控制中,进行人性化设计,变"人适应车"为"车适应人",最终实现车辆的安全、智能驾驶。文中确定了驾驶员特性辨识系统,以转向行为为例,对驾驶员特性进行分类,并采用神经网络方法建立了辨识模型,在驾驶模拟器上对所研究方法进行实验验证。结果表明,所建立的驾驶员特性辨识模型有较高精度,能对驾驶员特性进行预测,方法可行。

X-by-wire vehicle has more control freedoms than conventional vehicle,in which human（driver） factor can be incorporated into the integrated control of vehicle in designing control algorithms,i. e. pursuing humanized design and preferring ＂car adapting to driver＂rather than ＂driver adapting to car＂to achieve the safe and smart driving of vehicle. In this paper,the identification system of driver characteristics is defined,and with steering behavior as example,driver characteristics are classified. The identification model is set up by using neural network method,and the experimental verification of the methods in question is conducted on driving simulator. The results show that the identification model for driver characteristics is of high accuracy and the method is feasible,being able to well predict driver＇s characteristics.