Stable Control Algorithm for Slope Movement of Differential Drive Robot in Non-Flat Terrain

In order to ensure stable movement of differential drive robots on slopes, a control algorithm for stable movement of differential drive robots on slopes in non-flat terrain is proposed. By planning the wheel motion conversion process that adapts to changes in terrain slope, the differential drive robot generates wheel motion sequences that are suitable for different slopes. Construct an evaluation function based on stability constraints, select a state with good motion as the next motion state of the robot, and adjust the lateral position of the robot to meet the stability margin requirements. On this basis, a deep enhancement method based on RBF neural network and Q-learning is introduced. With the change of the current environmental state value of the differential drive robot, through continuous training and learning, the network weights and wheel motion parameters are updated in real time, and the optimal execution wheel motion mode is selected to achieve stable slope movement control of the differential drive robot in non-flat terrain. The experimental results show that under non-flat ground conditions, the proposed algorithm has small changes in the horizontal velocity of the center of mass, the energy curve is consistent with the ideal, the stability margin is greater than 0.60 units, and the success rate of wheel motion conversion is greater than 92%; On a gentle slope, with a stability margin of 0.75, a moving speed of 1.0[Formula: see text]m/s, a wheel motion conversion success rate of 98%, a recovery time of 2.0[Formula: see text]s, and an error of 0.1[Formula: see text]m, the differential drive robot can achieve stable movement in non-flat terrain.