A Reinforcement Learning based controller for optimal speed control of a DC motor using deep Q-network algorithm

Reinforcement Learning (RL) has been gaining significant attention in recent years as a powerful tool for solving complex control problems. This paper presents a possible application of an RL algorithm called deep Q-network (DQN) to the speed-tracking control of a DC motor. The proposed approach opens up new possibilities for the application of RL in the control of DC motors and other dynamic systems since it proposes a direct RL controller able to drive the motor in a realistic and fully randomized scenario.