基于驾驶员意图识别电动汽车平路起步控制策略研究

Study on Control Strategy for Starting Up On Flat Road of Electric Vehicle Based on Driver Intention Recognition

为改善纯电动汽车平路起步的操控性,提出了一种基于驾驶员意图识别的电动汽车平路起步控制策略.将驾驶员意图分为缓慢起步,急速起步,正常起步,利用k-medoids聚类算法对数据进行离散化处理,利用双层隐形马尔科夫模型对驾驶员的起步操作意图进行识别,分别制定了控制策略,使得缓慢起步状态下能将车辆稳定在较低车速,急速起步状态下能够最大利用车辆的动力性.仿真结果表明,该控制策略能正确识别驾驶员的起步操作意图,在缓慢起步状态下能够将车速稳定在5km/h,急速起步状态下能够达到地面能提供的最大加速度,保证了车辆起步时的动力性、安全性、操控性,实现既定的控制要求.

A control strategy for starting up of electric vehicle based on drtver retention was propose to improve starting handling. Driver intention can be divided into slow start, rapid start, and normal start, k-medoids clustering algorithm was used to discretize the data. Double hidden markov model was used to identify driver＇s operation intention. Make control strategy for three models. Slow start mode should keep vehicle at low speed, Rapid start state can make use o{ the vehicle＇s power performance maxi- mally. Simulation results show that the control strategies identify the driver＇s intentions correctly. Slow start mode can keep vehicle at 5 km/h. Rapid start can reach the maximum acceleration. Control strategy can ensure the vehicle started power performance, safety, handling, and achieve the established control requirement.