基于PCA-SVM的DCT汽车驾驶员起步意图识别

Identification of DCT Vehicle Driver’s Starting Intention Based on PCA-SVM

针对目前对于双离合自动变速器(Dual Clutch Transmission,DCT)汽车驾驶员起步过程中意图变化的研究有限,且现有方法的识别准确率不高这一问题,提出了一种基于主成分分析(Principal Component Analysis,PCA)和支持向量机(Support Vector Machines,SVM)的变化起步意图识别模型。根据驾驶员针对车辆反馈的反应时间,将起步过程划分为6个阶段;基于K均值聚类确定各个阶段的缓慢起步、一般起步、紧急起步这3种起步意图的界限;基于主成分分析挖掘出各个阶段起步意图识别的新特征;在此基础上构建6个SVM模型,并利用这6个模型分别对各阶段的起步意图进行识别。经过验证,该模型的平均测试准确率为94.92%,比只利用线性SVM模型高16.89%,且单个模型的平均耗时为0.008 s,能够快速有效地识别出DCT汽车驾驶员的起步意图。

Given the limited research on the driving intention change of DCT car drivers during the starting process,the paper proposed a driver intention recognition model based on principal component analysis(PCA)and support vector machine(SVM)to improve the recognition accuracy.The starting process was divided into six stages according to the driver’s reaction times to the vehicle’s responses.The boundaries of slow start,normal start and emergency start were determined based on K-means clustering.By using principal component analysis,six SVM models were constructed to identify the starting intention at each stage.The results show that the average test accuracy of the model is 94.92%,16.89%higher than that of the linear SVM.The average time spent by a single model is 0.008 s and therefore the starting intention of the DCT car drivercan be quickly and effectively identified.