基于人工神经网络的汽车闭环系统的研究

摘要

Presents a new approach to model tire and driver by establishing the nonlinear mapping relationship between input and output using artificial neural networks, the neural networks based tire model and driver model and corresponding vehicle model used for simulation of the closed loop system, and concludes from experimental results that neural networks can be used to generate tire side slip characteristics model and driver model with reasonable accuracy and high computational efficiency.%对于驾驶员-轮胎-汽车闭环系统的研究是全面评价和预估汽车操纵稳定性能的一条有效途径，但用一般数学方法难以精确描述轮胎和驾驶员的非线性特性.利用人工神经网络反映轮胎和驾驶员输入、输出特性之间的非线性映射关系，并将它们与相应的汽车模型结合进行闭环系统仿真.结果表明，人工神经网络可以产生出具有相当准确性及很高计算效率的驾驶员模型和轮胎侧偏特性模型.