Support Vector Machine Based Assessment System on Shift Quality for Vehicles: Theory, Structure and Application

Support vector machine (SVM) is a new statistic method which could make a good prediction with limited instances. Compared with artificial neutral network (ANN), SVM can provide better genetic ability. Shift quality is one of important vehicle performances, because it can be" felt" directly by the terminal customers. Therefore, how to make a correct assessment during the development is an important duty for vehicles manufacturing. The study provides an assessment system on shift quality based on SVM method. In order to verify the ability of the new method, the model trained by one automated manual transmission (AMT) car was applied on some other AMT vehicles, and the predicted results were compared with subjective rating results by expert drivers and analyzed to identify the potential of this new assessment system.