CPG-MPC controller for wheel-fin-flipper integrated amphibious robot
Yue Qiao,Wang Wei,Yunxiang Li,Shengzui Xu,Lang Wei,Xu Hao,Re Xia
DOI: https://doi.org/10.1108/ir-01-2023-0008
2023-09-29
Industrial Robot
Abstract:Purpose The purpose of this paper is to introduce a motion control method for WFF-AmphiRobot, which can effectively realize the flexible motion of the robot on land, underwater and in the transition zone between land and water. Design/methodology/approach Based on the dynamics model, the authors selected the appropriate state variables to construct the state space model of the robot and estimated the feedback state of the robot through the maximum a posteriori probability estimation. The nonlinear predictive model controller of the robot is constructed by local linearization of the model to perform closed-loop control on the overall motion of the robot. For the control problem of the terminal trajectory, using the neural rhythmic movement theory in bionics to construct a robot central pattern generator (CPG) for real-time generation of terminal trajectory. Findings In this paper, the motion state of WFF-AmphiRobot is estimated, and a model-based overall motion controller for the robot and an end-effector controller based on neural rhythm control are constructed. The effectiveness of the controller and motion control algorithm is verified by simulation and physical prototype motion experiments on land and underwater, and the robot can ideally complete the desired behavior. Originality/value The paper designed a controller for WFF-AmphiRobot. First, when constructing the robot state estimator in this paper, the robot dynamics model is introduced as the a priori estimation model, and the error compensation of the a priori model is performed by the method of maximum a posteriori probability estimation, which improves the accuracy of the state estimator. Second, for the underwater oscillation motion characteristics of the flipper, the Hopf oscillator is used as the basis, and the flipper fluctuation equation is modified and improved by the CPG signal is adapted to the flipper oscillation demand. The controller effectively controls the position error and heading angle error within the desired range during the movement of the WFF-AmphiRobot.