论文部分内容阅读
为模拟驾驶员在跟踪期望路径过程中的驾驶决策过程,论文基于目标位置的思想对驾驶员的学习过程进行了建模。在建模过程中,利用三次B样条曲线灵活性的特点生成车辆对于目标位置的换道路径,同时,借助模糊神经系统模拟驾驶员跟踪目标位置过程的决策机制。并且提出了一种新的学习算法,即通过将模糊论域划分为若干局部线性网络以此拟合非线性函数。利用该算法能模仿驾驶员驾驶过程中的自学习能力。最后,对车辆躲避障碍物进行了仿真,结果表明该模型能较好的模拟驾驶员在换道过程中的驾驶行为及其自学习过程。
To simulate the driver’s decision-making process in tracking the desired path, the thesis models the driver’s learning process based on the idea of target location. In the process of modeling, the lane-changing path of the vehicle to the target position is generated by the characteristic of cubic B-spline curve flexibility. At the same time, the fuzzy neural system is used to simulate the decision-making mechanism of the driver tracking the target position process. And a new learning algorithm is proposed, which is to fit the nonlinear function by dividing the fuzzy universe into several local linear networks. The algorithm can simulate the driver’s self-learning ability during driving. Finally, the simulation of the vehicle avoiding obstacles is carried out. The results show that the model can better simulate the driver’s driving behavior and its self-learning process during lane changing.