A multiple-step vehicle torque demand forecasting problem is investigated in this paper. Neural network method is applied to forecast toque demand in the future five steps based on the measurements of vehicle speed and torque demand in the past three ...
A multiple-step vehicle torque demand forecasting problem is investigated in this paper. Neural network method is applied to forecast toque demand in the future five steps based on the measurements of vehicle speed and torque demand in the past three steps. Neural networks for forecasting are trained based on history drive cycles’ data and used to do multiple-step forecasting in another drive cycle to verify the performance. Results of the application have shown that the desired goal has been successfully achieved.