To ensure that a robot can maintain a stable contact force when it contacts the environment,an adaptive PID control for robot constant force tracking was designed based on a dueling deep Q-network.The contact process between a robot and the environment was analyzed,and a robot force controller based on a PID algorithm was constructed.The adaptive PID algorithm based on a dueling deep Q-network was proposed to adapt to changes in the external environment.In this algorithm,the dueling deep Q network was used to learn and find optimal control parameters.Finally,robot constant force tracking experiments were expanded on Coopeliasim and MATLAB software platforms.The simulation results show that the adaptive PID algorithm based on the dueling deep Q-network can achieve good force-tracking effects,verifying the algorithm′s feasibility;compared with a deep Q network algorithm,the average absolute value of the force error is reduced by 51.6%,and the convergence speed is improved,allowing the robot to track the external environment better.