Electric vehicle torque optimization method based on data driven predictive control

The invention provides an electric vehicle torque optimization method based on data driven predictive control, belongs to the technical field of electric vehicles, and aims to effectively perform optimized distribution of driving and braking torque of an electric vehicle so as to realize a tracking control method for the longitudinal velocity of the vehicle. The method comprises the steps as follows: firstly, appropriate excitation data are designed according to dynamic characteristics of a system, so that sufficient excitation for the system is guaranteed; secondly, a predictive output equation of the system is constructed by input/ output data obtained by excitation; then actuator rigid constraints of a motor, a battery pack and a brake are considered, and a cost function for torque optimization control is constructed with a model predictive control algorithm; finally, an optimization problem corresponding to the cost function is solved to obtain control input which then acts on the system, so that the system is controlled. For the control algorithm, all that is required is off-line simulation experiments when the excitation data are acquired, so that the development cost is lower.