Research On Handling Comfort Of Vehicle Gear Shift Lever Based On BP Neural Network And Evidence Theory

The vehicle gear shift lever is one of the most frequently used control devices in man-machine system, and it gets closely related to the handling comfort, the safety, and the health of driver. In the current research, evaluations on the handling comfort of vehicle gear shift lever are lack of combination of subjective and objective factors, because most of them mainly focuses on changing regularity of certain control force mechanical characteristic parameters or subjective feeling of driver. Therefore, this paper combines with the actual manipulative situation of vehicle gear shift lever to comprehensively analyze the key factors affecting the handling comfort. And the evaluation model for the handling comfort of vehicle gear shift lever is built based on neural network and evidence theory. The specific research work is shown as follows:1. Through the survey in car users, data analysis, as well as comprehensive analysis of related reference literature of domestic and foreign, this paper takes "personnel-device" as the breakthrough point to analyze affecting factors of the handling comfort of vehicle gear shift lever in the process of man-machine interaction.2. On the basis of the in-depth analysis of vehicle gear shift lever handling process and in accordance with the key factors affecting handling comfort, evaluation indicators system of the handling comfort of vehicle gear shift lever is built according to the theories of system evaluation.3. Through the study on BP neural network, the data synthesis methods of D-S evidence theory that usually implicated in the field of system evaluation, find that the output of a single BP neural network is instable, and sometimes even can influence the determination of evaluation results; and the data synthesis methods of D-S evidence theory can weaken this kind of instability, but its basic probability assignment is difficult to distribution. Therefore, this paper combines BP neural network with D-S evidence theory to establish the evaluation model for handling comfort of vehicle gear shift lever.4. Finally,5 real cars are analyzed to test the evaluation model, and the result shows that the model is feasible and effective in the evaluation of the handling comfort of vehicle gear shifts. At last, combining the evaluation results to discuss and give suggestions about designing and optimization for vehicle gear shift lever.