Research On Multi-gait Control Of Hexapod Robot Based On Fast Synchronization Strategy Of Hopf Oscillator

When robots operate in some unstructured terrain and high-risk situations,they often need to complete tasks independently,such as applications in mine detection,fire prevention and disaster relief.Legged robots not only have flexible motion forms,but also show strong adaptability in complex environments.Therefore,they have gradually become a research hotspot in the field of bionic robots.Gait research on hexapod robots is the basis for the development of legged robot technology,and the method based on central pattern generator(CPG)has become one of the main means of gait control.However,in order to make the hexapod robot quickly adapt to unexpected situations during walking,this method still needs further research.There is still room for improvement in terms of output form,gait switching,and fast self-adaptation.This paper makes further research on these issues.Firstly,inspired by the structural characteristics of ants,a bionic hexapod robot experiment platform was designed,a single-leg D-H model was established,and the forward and reverse kinematics solution process was derived.In addition,the robot-related hardware and software systems are introduced.Next,various oscillator models are compared and analyzed.Hopf oscillator is selected as the control unit of the CPG neural network,and the conditions of synchronous oscillation are analyzed from the perspective of phase dynamics.In addition,considering the robot will have differences in the actual movement process,through the simulation analysis of four frequency learning mechanisms,it is determined that this paper integrates the dual-rate error integral learning mechanism(DIL)into the CPG neural network,and builds a control architecture based on adaptive frequency adjustment module,which can not only realize the rapid adaptation of the frequency,but also keep the learned expectation value continuously oscillating.Then,the typical gait of the hexapod robot is planned,and the CPG rhythm signal is mapped to the input of the foot trajectory generator through the neural Q-Bézire formula,and then use the DIL mechanism for fast tracking and learning of the trajectory.Through the simulation analysis of each gait with Matlab,aiming at the problem of waveform jitter in the switching process,a method of continuously changing parameters is proposed,which realizes the fast and smooth switching between multiple gaits of the robot.Finally,a Gazebo simulation environment is built to simulate and analyze the fast synchronization learning strategy of frequency,which verifies the effectiveness of the method.The fast tracking and learning of the foot trajectory and the generation and switching of each gait are experimented.The correctness of the gait planning method in this paper is verified by observing the changes of joint angle and attitude angle during the movement.