基于深度强化学习-PI控制的机电作动器控制策略

Control strategy of electro-mechanical actuator based on deep reinforcement learning-PI control

为提高机电作动器在指令控制下的控制精度与跟踪速度,提出了一种基于深度强化学习的机电作动器指令控制方法。首先,根据机电作动器的结构,建立机电作动器的数学模型。然后,将深度强化学习与比例积分控制相结合,通过强化学习智能体与机电作动器系统进行在线交互从而获得奖励信号,并利用奖励信号实时整定PI控制器的参数,以实现对机电作动器的指令控制。最后,验证所提出的方法在机电作动器的指令控制下的效果。仿真得出的结果稳态误差更小,响应速度更快,验证了算法在机电作动器控制上的有效性。

In order to improve the control accuracy and tracking speed of electromechanical actuators under command control,this paper proposes an electromechanical actuator command control method based on deep reinforcement learning.First,according to the structure of the electromechanical actuator,the mathematical model of the electromechanical actuator is established.Then,by combining the deep reinforcement learning with proportion-integration(PI)control,the reinforcement learning agent interacts with the electromechanical actuator system online to obtain the reward signal.Using the reward signal to tune the parameters of the PI controller in real time to realize the control of the electromechanical actuator.Finally,the effectiveness of the proposed method is verified under the command control of electromechanical actuators.The simulation results show that the steady-state error is smaller and the response speed is faster,which verifies effectiveness of the algorithm in the control of electromechanical actuators.