Double-layer CPG Motion Planning And Design Pneumatic Muscle Spine

Flexibility,stability,and bionicity are the major points of studying quadruped robot movement.Spine as a link between the front and rear limbs lacks minute study about its trajectory in current gait planning.In this paper,based on bionicity and flexibility of spine,we research on central pattern generators(CPGs)neural network model.Through the establishment of double-layer CPG mathematical model to plan the trajectory of spinal segment,which is applied to designed pneumatic muscle parallel spine festival.Firstly,from the perspective of biology and bionics,this paper analyzes the structure of CPG network and its intelligent characteristics.Using improved hopfield oscillator model to model CPG neural network,and propose a double-layer CPG trajectory generation method to simulates biological motion planning mechanism.This includes rhythm layer that determines the phase of joint motion for planning different gait,and pattern layer that determines the form of joint motion is used to plan muscle flexion and extension information.Calculate and analyze the parameters of model and output curve of each layer to verify the ability of stability,anti-interference and fast response of model to input signal.Secondly,according to yaw and pitch movements of cheetah's spine in trot and gallop gait,observe the effect of movement mode changes on the movement trajectory,and design double-layer CPG closed-loop system witch self-adaptive external sensory information.The working form of neural-muscle system is used to construct double-layer CPG model,output a fitted mapping relationship with the target angle of spine,and arc tangent motion equation between angle and muscle drive.Build double-layer CPG system in Simulink,analyze simulation results of spinal with two degrees of freedom through different information inputs.Then,design a parallel spine virtual prototype in Webots simulation environment,carry out co-simulating of Webots and MATLAB.Verify the rationality of double-layer CPG in planning spinal motion,the error analysis is performed between actual output curve and ideal curve.The results show that double-layer CPG can simulate the biological movement coordination mechanism and can complete mutual transformation of yaw and pitch motion states according to input information.This algorithm has fast response speed and can achieve ideal output in 0.1s.The error is less than 1.5 °,which can achieve complete tracking.Finally,design structure of parallel vertebral segment according to the results of theoretical research and simulation analysis,build mechanical object and controller object.Add PID control algorithm on double-layer CPG planning strategy to verify the correctness of method.The results show that double-layer CPG is suitable for spinal vertebral motion,and can change the motion rhythm and motion mode according to input information in real time,which reflects the coordination relationship between input and motion response.