Proposed Shift Quality Metrics and Experimentation on AMT Shift Quality Evaluation

A method to evaluate AMT shift quality based on neural network is implemented in Matlab/Simulink and an objective evaluation software is designed specially, which enables the possibility to produce ratings. The neural network is trained by data swatch attained after selecting the quality shift metrics. Tests show that it is able to produce grades to evaluate AMT shift quality. And the correlation between the developed evaluation method and subjective ratings given by drivers is proved.