Shift quality evaluation system based on neural network for DCT vehicles

A vehicle simulation model for Dual Clutch Transmission (DCT) is designed in Matlab/Simulink, and more importantly, an objective description of the method to evaluate its shift quality based on neural network (NN) is implemented. The objective evaluation method adopted the theory of NN prediction; NN is trained and tested with sample vectors comprised by shift quality evaluation index, subjective evaluation rates. The correctness of the vehicle control algorithms and algorithms extracted evaluation index are validated by simulation, and the per-development performance of DCT is predicted successfully. The experimental results show that the evaluation system cannot only be used to realize the objective evaluation of shift quality, but also be used as a reference and guidance role for the calibration vehicles equipped with dual clutch transmission.