一种优化参数的支持向量机驾驶意图识别

摘要

With the accelerator pedal opening,and its change rate as input parameters,and acceleration intentions are divided into slow acceleration,general acceleration and rapid acceleration,this paper sets up an electric vehicle driving intention recognition model based on support vector machine.In order to solve the problem that the search range is uncertain as optimizing the parameters of support vector machine based on particle swarm optimization,this paper proposes an adaptive particle swarm optimization algorithm.At first,the optimal range of the parameters is obtained by the grid search method,and then the particle swarm algorithm is optimized in this range.Finally,a higher accuracy of the classification results and shortened training time are obtained.The experimental results show that the prediction model established by this adaptive particle swarm optimization algorithm for support vector machines is highly accurate and reliable,and provides a new method for the identification of driving intention.The results of the driving intention recognition can be used for the research of the following pure electric vehicle drive control strategy,and further improve the driving performance of the vehicle.%以加速踏板开度、加速踏板开度变化率为输入参数,将加速意图分为缓加速、一般加速和急加速,建立了基于支持向量机的电动汽车驾驶意图识别模型.为了解决粒子群算法优化支持向量机参数时寻优范围的不确定性,导致搜索效率不稳定的问题,提出了一种自适应粒子群算法:先通过网格搜索法确定出粒子群算法参数寻优的最佳范围,再由粒子群算法在此范围精确寻优,最后得到了更高准确率的分类结果和缩短了的训练时间.通过仿真实验验证,运用这种自适应粒子群优化支持向量机建立的预测模型辨识度高,模型准确可靠,为驾驶意图的识别提供了新的方法.驾驶意图识别的结果可用于后续的纯电动汽车驱动控制策略的研究,进一步提高汽车的驾驶性能.