Research On Conrtol Strategies For Chaotic Gait Of Passive Biped Robot

Achieving stable,efficient and natural human-like gait is one of the important research goals in the field of biped robot.Compared with traditional active robots,passive biped robots have lower energy consumption and a highly human-like walking gait,which has become a hotspots in the research field of biped robots.However,because the passive biped robot has complex dynamic characteristics and is extremely sensitive to the external environment,it might show unstable non-periodic gait such as period-doubling bifurcation and chaos gait when certain parameters are changed.When the biped have a non-periodic gait,the stability of the robot system deteriorates and even falls.This is very unfavorable for robots.Therefore,it is necessary to control the biped robot from the unstable gait to a stable periodic gaitIn this paper,a passive biped robot is taken as the research object,and the effective chaos control strategies are proposed for the chaotic gait caused by the slope change,which provides a strong theoretical support for the stable walking of the biped robot under actively controlling.The main research work of this paper is as follows:Firstly,based on the human gait division,the hybrid dynamic model of the biped passive robot is established by using the Lagrange method and the conservation of angular momentum.On this basis,the Poincaré mapping method is applied to analyze the progressive stability of robot gait and the Newton-Raphson iteration method is applied to solve the gait fixed point of robot.For the non-periodic gait of robot,the slope angle is regarded as variable,and then Floquet multiplier and gait bifurcation principle are used to analyze change regulation of the biped gait under the change of slope angle.Secondly,in view of the highly nonlinear and hybrid dynamics of biped robots,a gait prediction method based on dynamic fuzzy neural network is proposed.The identifiable conditions of the robot hybrid system is analyzed and deduced from the asymptotic stability,and the dynamic fuzzy neural network is used to establish the prediction model of the robot's swing stage and collision stage respectively.According to the coupling andconversion of the same stage in the two stages,the two neural networks are cascaded.As a result,the accurate prediction of the biped's gait is achieved,which provides strong support for the control method based on gait prediction.Once again,combining with the accurate prediction of the gait of biped robot,a DFNN-based predictive feedback control method for chaos gait is proposed.Using the robot's current state information and the predicted future state,an adaptive predictive feedback controller is designed to achieve effective control of the chaotic gait of the robot.Finally,aiming at the problem that the fixed point of the biped robot is difficult to obtain quickly,a fast fixed point acquisition method based on DFNN and Newton-Raphson iteration is proposed.Then based on the basic idea of OGY method and feedback control,a hybrid control method based on adaptive energy feedback and OGY method is proposed to make up for the shortcomings of the traditional OGY method such as real-time difference in the chaotic gait control of biped robots.Through the simulation,the effectiveness of the proposed control method is verified.