A Fuzzy Neural Network-Based Robust Control Algorithm for the Trajectory of a Roadway Inspection Robot

In order to meet the requirements of complex tunnel inspection and ensure the reliable operation of inspection robots under external interference and terrain changes, a trajectory robust control algorithm based on a fuzzy neural network is proposed. This algorithm combines the motion characteristics of the inspection robot to analyse the motion variables that affect the trajectory. The deviation and deviation rate of the left and right track running speed and steering angle are used as control variables and input into the PID controller. By adjusting PID parameters online through fuzzy neural networks, a PID robust controller is constructed to achieve trajectory control. Tests have shown that the controller performs best when the PID control parameters are set to 0.25, 0.65 and 0.55. The steering angle robustness margin is less than 2° in both single and complex scenarios, and the maximum trajectory deviation is only (0.3,0) cm, effectively achieving robust trajectory control of the tunnel inspection robot.