Hybrid electric vehicle energy management method based on reinforcement learning

Abstract

The invention relates to a hybrid electric vehicle energy management method based on reinforcement learning. The method comprises the following steps: obtaining energy management strategies of a hybrid electric vehicle under different cycle conditions based on a Q-learning algorithm in reinforcement learning; writing the energy management strategy into a microcontroller; determining a current cycle working condition, acquiring current driving parameters and transmitting the current driving parameters to the microcontroller by the data acquisition system, acquiring a control action by the microcontroller based on an energy management strategy under the current cycle working condition, and transmitting the control action to the vehicle control unit; and adjusting the power system by the vehicle control unit according to the control action. Compared with the prior art, the energy management strategies of the hybrid electric vehicle under different cycle working conditions are obtained based on reinforcement learning and written into the microcontroller of the hybrid electric vehicle, the optimal control action under the current state can be quickly found only by looking up the table when the vehicle runs, the speed is high, And the sampling frequency of state monitoring of the hybrid electric vehicle at present or even in the future can be met.

Description

technical field [0001] The invention relates to the technical field of hybrid electric vehicle control, in particular to an online energy management method for a hybrid electric vehicle based on reinforcement learning. Background technique [0002] In order to save resources, reduce environmental pollution, and achieve energy saving and emission reduction, hybrid electric vehicles have become one of the important directions for the development of the automobile industry today. As a key control technology for hybrid electric vehicles, energy management strategies directly affect the fuel economy of automobiles and become Research focus of hybrid power system. [0003] In recent years, the research on energy management strategies of HEVs can be mainly divided into two categories. One is rule-based control algorithms, such as logic threshold and fuzzy logic control algorithms. Rule-based control algorithms have clear logic and fast calculation, but the optimization effect is l...

AI Technical Summary

Problems solved by technology

However, this method is an energy management strategy that collects data in real time and optimizes it in real time. It has the common problems of many energy management strategies: a large amount of calculation, a long iteration time, and it is difficult to take into account real-time and optimal solutions. Sampling frequency, or reducing the number of iterations, the optimization effect will obviously decrease, and it is difficult to apply it on a real vehicle

Images