Research On Motion Control Of Quadruped Robot Based On CPG In Complex Terrain

Quadruped robots have great potential value of application in military,transportation,survey and other fields because of its strong adaptability to different environment.But most of the existing quadruped robots can only walk on the flat ground in the laboratory,and are difficult to adapt to the complex environment.Aiming at the problem of motion control of quadruped robots in complex terrain,this paper starts with bionics,explores the method of biological motion control,and designs a control algorithm based on central pattern generator(CPG),which can be applied to incline,step and trench.In order to solve the problem that robot lacks the ability to sense the environment,this paper designs a new type of foot with the ability of contact detection.The foot can detect the contact between the robot's foot and the ground in different directions,and then judge the terrain change.In this paper,the influence of internal parameters on the output signal of Matsuoka oscillator is determined by the linearization method.By introducing transient interference signal and gait matrix to control the robot's gait,the problem of slow response of the robot in gait conversion is solved.Then,the paper designs the trajectory of the robot's foot on the flat ground,and then extends the trajectory to the inclined plane motion combined with the fuselage attitude adjustment algorithm.For the complex terrain,this paper distinguishes the terrain that the robot may face according to different foot contact signals,and designs corresponding foot trajectory planning algorithms.In order to solve the problem of poor continuity when the robot moves on a continuously changing slope,this paper proposes a terrain prediction algorithm based on foot contact detection.Through the recorded information of the historical landing point,an arc model is used to predict the change trend of the terrain,and then the foot trajectory of the next cycle is planned in advance to improve the motion continuity of the robot.In view of the instability phenomenon that may occur in robot motion,this paper proposes an attitude adjustment algorithm based on model predictive control(MPC).By adjusting the leg length to control the robot's posture,it can prevent the robot from falling due to the tilt of the fuselage.Through the co-simulation of MATLAB / Simulink and v-rep software,and the physical experiment of Fly Dog quadruped robot,the validity of the robot control algorithm designed in this paper is proved.