Design Of Control System For Walking Robot Based On Inverted Pendulum Of Reaction Wheel

After a long natural selection,the four-legged mammal gradually evolved into a limb that can adapt to complex environments.Ordinary tracked,wheeled cars or robots made by humans are difficult to drive in complex terrain.The bionic quadruped robot manufactured according to the principle of bionics has great potential in complex terrain reconnaissance and field material t ransportation,and has become a hot spot in robot research.The key to maintaining the lateral balance of the body during walking or standing is the lateral support force provided by the left and right legs.In this case,the quadruped robot needs to add four degrees of freedom to maintain the lateral balance.This makes its control algorithm more complicated.In comparison,the system uses only one reaction wheel motor to maintain lateral stability.In this paper,the mammal in nature is used as a template,and the four-legged walking robot in the inclined state is equivalent to the inverted pendulum model of the reaction wheel.By studying the dynamic model of the inverted pendulum of the reaction wheel,the optimal control is realized,so that it can react quickly during the tilting process.And the amount of overshoot is small.Aiming at the characteristics of unstable,nonlinear and strong coupling of the inverted pendulum model,the fuzzy adaptive tuning PID algorithm is used to solve the shortcomings of the traditional PID algorithm and can not adapt to the nonl inear system.For the traditional basic PID parameters,manual tuning is needed,but the artificial experience method can not find the most suitable parameters of the system.This paper designs a cost function that is more suitable for the system,and uses the strategy gradient reinforcement learning algorithm to find the global or Local optimal solution.After deriving the dynamic model of the four-legged walking robot,this paper establishes the link coordinate system of each leg of the four-legged walking robot,and solves the link motion equation to control the leg motion based on this coordinate system.The reaction wheel inverted pendulum is mounted in th e quadruped walking robot to maintain the lateral stability o f the quadruped robot.Finally,the above algorithm is applied to control the reaction wheel inverted pendulum so that the reaction wheel inverted pendulum remains stabl e for a long time,and the four-legged robot with the reaction wheel inverted pendulum can maintain lateral stability.