Research On Free Gait Planning And Motion Control Of Hexapod Robot

As a kind of high redundancy and multi degree of freedom foot robot,hexapod robot has incomparable advantages in adaptability,reliability,motion performance and so on.However,the over complex nonlinear structure brings new challenges to the motion control and gait planning of hexapod robot.In order to ensure the motion control accuracy of hexapod robot,improve the adaptability of hexapod robot to the environment,and then realize the autonomous walking of hexapod robot,the motion planning and control of hexapod robot has gradually become a key problem in the research field of hexapod robot in recent ten years.Firstly,this paper analyzes the research status of hexapod robot at home and abroad,and summarizes the shortcomings of the existing research from two aspects of control system and gait planning.The mathematical modeling of the whole structure of the hexapod robot is completed,and the kinematics and dynamics are analyzed;Based on the traditional planning of foot trajectory and body trajectory,the multi foot cooperative control of hexapod robot is studied,and the coupling relationship between the two trajectory planning is analyzed,which provides a prerequisite for robot motion control and gait planning;In this paper,the artificial neuron is simulated and analyzed,the chaotic characteristics of neuron are explored,the topology of traditional central nervous model generator is modified,the chaotic neuron is added,the chaotic neural network model is proposed innovatively,and the corresponding neural network functional unit is designed to convert the complex and irregular sensor signal into the standard and processable input signal.Finally,the vibration rhythm signal which can control the joints of the hexapod robot is generated,and the neural network model of motion control of the hexapod robot is established.The gait control method based on CPG(Center Pattern Generator)principle is studied.The topology of gait control is improved,the amount of joint control is reduced,and the efficiency of gait control is improved;Taking the discrete landing point of hexapod robot as the research object,a discrete gait model is established,which transforms the complex free gait planning problem into a mathematical sequence scheduling problem;This paper analyzes the problems of traditional deep reinforcement learning algorithm,optimizes the experience playback mechanism by using sampling algorithm,and proposes IPER algorithm(Improved Prioritized Experience Replay),which is verified by co-simulation platform,The results verify the feasibility and accuracy of the algorithm.Finally,the hardware and software system of the motion planning and control system of the hexapod robot is designed.CRIO is the core controller of the hardware system,and the circuits of pose information acquisition module,foot information acquisition module and remote communication module are designed.The software system takes Lab VIEW software mixed programming as the core,mainly including communication system,human-computer interaction system and so on.The experimental platform of hexapod robot is verified,and the unstructured terrain movement experiment and communication interruption experiment are completed.The feasibility and stability of the control system of hexapod robot are verified.