Prediction method of target torque for Pure Electric Vehicle based on BP neural network

New European Driving Cycle (NEDC) condition test for a Pure Electric Vehicle (PEV) has been performed on a rotary drum test bench. Then with the help of the test data, the prediction model of vehicle target torque is established based on Back Propagation (BP) neural network, which aims at accurately predicting the target torque of driving conditions. Finally, use the other test data to validate the model. Simulation results confirm that this model can accurately predict the change trend of target torque, but a few errors exist between predicted value and actual value. Therefore, it is necessary to take measures to improve the model for a high precision of prediction.