Gait Control Strategy Of Biped Robot With Multi-degree Of Freedom

The conception of biped robot comes from human beings. Development up to now, biped robot has been the integration of many technologies, such as electronics, mechanics, computer, sensor, artificial intelligence and automation, which is the important symbol to measure the overall level of national science and technology. Of all human behaviors, walking function is one of the most characteristic behaviors possessed by biped robot. The mechanism is a very complex nonlinear system with strong coupling and multi-degree of freedom, which designed on the basis of human body. So, the control of walking motion is the emphasis and difficulty of research in biped robot. Based on this background, this paper carried out research for the ideal gait dynamic model, complex gait planning method, many artificially specified parameters and huge computation in current walking control methods.In this paper, a walking control system of biped robot based Kinect was proposed to realize anthropomorphic walking motion naturally and steady. The main contents include:(1) build local coordinate system of human joints via fitting human motion data captured by Kinect based on the least-squares fitting method. Build human body kinematics model with homogeneous coordinate transformation method to resolve human posture.(2) in terms of gait planning, determine the mapping joints relationship between human body and biped robot, generate the angles of robot joint with human body joint information. Build robot model by matching kinematics model with human body based on D-H.(3) in terms of gait control, firstly, derive ZMP expression via analyzing kinetic characteristic of biped robot to build single-particle model. Then, build stable walking area by FSR sensor and analyze robot kinestate with the sum of angles. Last, seek optimal ankle joint angle by gradient descent method to control gait of biped robot based on ZMP stability criterion.(4) according to ZMP situation, analyze the stability of robot walking motion and evaluate the stability by safety factor.(5) use Webots emulator to verify the walking control system of biped robot with Nao robot.