Reaserch On Theory For Dynamic Walking Gait Planning Of Biped Robot

Along with the development of high technology and the deepening of the research on artificial intelligence, the humanoid robot will come into people's life gradually. Now the key problems humanoid robot faced are robot artificial intelligent and motion control. Walking movement control relies on an effective sample of movement. This paper mainly studies the dynamic walk of robot and aim to provide effective dynamic walk samples generating method for biped robot.Firstly, the paper brings in fully Cartesian method to the biped robot coordinate modeling. Without the use of Euler parameters, make program more simple and easier. And then provides the rigid body and hinge constraint equations, the calculation method of momentum and moment of momentum and the ZMP description in fully Cartesian coordinates. At last, analyses the relationship between the joint angle and Cartesian parameters.Secondly, by analyzing the relationship between the centroid trajectory based on the linear inverted pendulum model and walking parameters, provides a reference for movement parameter selection. Through the analysis of the relationship between ZMP and the robot walking trajectory, propose a new centroid trajectory planning method with mobile ZMP and fixed ZMP trajectory in single foot support period. This method improves the fast walking ability of robot. And then provides the centroid trajectory planning method in start step, middle step and stop step. And also gives the trajectory planning method of the swimming foot.Finally, based on the actual robot model, plans the trajectory of each joint. And proposes a walking samples generation method called walking unit sample combination method. This method provides a new idea for robot real-time walking samples generation. At last, the method is verified through ADAMS simulation.