摘要
针对目前汽车换挡平顺性评价方法的不足,提出了一个基于证据理论和模糊神经网络的汽车换挡平顺性评价方法。运用证据理论对不同驾驶员给出的主观评价进行数据融合;通过模糊神经网络对由仪器测得的客观评价指标和经证据合成后的相应主观评价构成的样本向量进行学习和训练,建立了汽车换挡平顺性评价系统计算和试验结果表明,该方法克服了主观评价和客观评价各自的缺点,能客观、准确、有效地评价汽车换挡平顺性。
In view of the shortcomings of existing methods for vehicle shift-feel evaluation, a new scheme for shift-feel evaluation based on evidence theory and adaptive neural fuzzy inference system （ANFIS） is proposed Firstly data fusion or evidence synthesis is conducted on subjective ratings of various drivers by applying evidence theory; then the sample vectors composed of the objective evaluation indicators measured by instruments and the corresponding subjective evaluation results after evidence synthesis are trained with ANFIS; finally a shift quality evaluation system is established. Computational and experimental results show that the proposed scheme overcomes the respective shortcomings of subjective and objective evaluations and can achieve vehicle shift-feel evaluation accurately, effectively and objectively.
出处
《汽车工程》
EI
CSCD
北大核心
2009年第4期308-312,共5页
Automotive Engineering
基金
江苏省汽车工程重点实验室开放基金项目(QC200603)
江苏省交通科学研究计划项目(06C04)资助。
