基于BP神经网络的汽车离合器舒适度评价方法研究

Research on the Comfort Evaluation of Automobile Clutches Based on the BP Network

离合器操作性能的舒适性评价,需要将主观评价因素与客观测试参数相结合。根据离合器参数对操作舒适性的影响分析,确定了7个参数测试的主要指标,搭建了进行盲测的实车模拟操作测试平台,利用基于B P神经网络(B P网络)的模型对实验数据建立输入-输出模式映射关系。基于样本的模型训练试验结果表明:模型的主客观评分无明显的系统性偏差,神经网络测试模型可较为客观地反映离合器的操作舒适性。这一方法也有望拓展到汽车行业中需要为主观舒适性提供客观数据的其他产品测试环节。

The comfort evaluation of the clutch operating performance needs to be implemented by combining the subjective evaluation factors with the objective test parameters. Based on the effect analysis of the clutch parameters on the operation comfort, seven evaluation factors are determined. An experimental platform for the blind simulation test of clutches is built.Based on the model of BP neural network, the mapping relationship between the input experimental data and the output result pattern is established. The results of model training test based on samples show that no systematic deviations exist in the subjective and objective scores of the model. The test neural network model can reveal the operation comfort of clutches objectively. This approach is also expected to be extended to other product testing in the automotive industry which need to provide objective data for the subjective comfort test.