Abstract:

In order to improve the energy recovery rate of electric unmanned vehicles during operation and to ensure the safety and economy of vehicle operation， a regenerative braking control strategy for electric unmanned vehicles based on speed prediction is proposed. The road condition detection algorithm based on offline training neural network model and the vehicle speed prediction method based on improved Markov chain are used to make the control process more accurate and stable. The sliding sampling window method is used for online pattern recognition of the vehicle driving state， and the predicted speed values are converted into the power demand for vehicle operation， and then the braking torque applied to the vehicle tires is solved by a model prediction controller to determine the optimal solution for the electric unmanned vehicle braking for different braking torques applied to each tire. The experimental results indicate that electric unmanned vehicles employing regenerative braking control strategies can effectively manage the efficiency of energy recovery during braking and extend their cruising range.

Key words: electric unmanned vehicles, regenerative braking control strategy, recurrent neural network, speed prediction