The Gait Planning And The Balance Control Of The Quadruped Walking Robot Based On ZMP

The quadruped walking robot has much better adaptation to environment in human living-space than wheel type, crawl type and pedrail type mobile robots. Researchers at home and abroad have been undertaken extensive research on the stability of quadruped walking robot. Three problems have been focused on in this paper:establishment of kinematics and dynamics models for the lower limbs of robot; gait planning and simulation of robot; design and simulation of robot fuzzy control system. The main tasks are as follows:Firstly, the degrees of freedom configuration and structure of the robot were designed; the direct-inverse kinematics models for robot based on homogeneous coordinates transformation theory was made in pole coordinates, the base for the following gait planning was provided. Dynamics models with the Lgrange dynamic equations for the single-foot support were established, which can be used in control.Secondly, the basic principle and method of the robot gait planning was analyzed, gait planning on robot static walk is based on barycenter imitation. The method of gait planning based on model, which concludes body gesture control, gait control and foot trail control is used. Considering that the rigidity of the quadruped walking robot's leg joint is not enough, and the foot may skid, causing that the kinematics planning trail of the foot trail control don't tally with the trail of the real movement. That can not make the quadruped walking robot walk as the designed movement way. Therefore, we introduce the ZMP trail planning method in a support polygon into the foot trail control, which to make sure that the quadruped walking robot can walk in a best gait by planning the movement gait of the foot relative to body, when the robot's foot is supporting.Lastly, on the procession of the dynamic balance of the quadruped walking robot, the most important is to keep the robot's balance by controlling the ZMP point into a stable area. The traditional method of figuring the ZMP point takes a lot of calculation, and the current position, velocity, and acceleration of the every part of the quadruped walking robot's must be known. In this paper, a method based on information of the distribution of the power on sole is put forward. We fix the FSR force sensor on the sole to get the data of the current ZMP point and velocity difference. The robot stable and ZMP fuzzy control was designed on fuzzy logical theory, which was imitated in MATLAB. The controller can control the ZMP to a expected position by amending the joint ankle of the robot, that can make the quadruped walking robot keep its balance in real time.