Research On The Analyssis Of Quadruped Robot Dynamic Stablity And Control

With the rapid development of robot technology,to be able to walk in the complex environment of increasing demand for special robot。Legged robot has the high adaptability to the environment, it can be applied in the fields of scientific investigation, national defense, aerospace exploration etc, in recent years, research on quadruped robots was most exciting topics in the field of robotics. It required the robot not only to adapt the regular and known environment, but also to adapt to irregular, unknown environments by future. Although the researchers work completely on the static quadruped walking, there is less study on the dynamical quadruped walking-especially on irregular terrain. Whereas the static quadruped walking can not be applied to rapid and flexible movement, the research of walking dynamically on irregular terrain for a quadruped robot is of great importance.In this paper, we carried out for the research on quadruped robots gait generation of parametric and stability control method based on animal Neural network control mechanism.Firstly analyzes quadruped animal legs joint movement rules, use this rule of the gait, imitation sample dog designed quadruped robot parametric gait generation algorithm. Using the van der Pol generated rhythm oscillator signal; Analysis of the mapping between gait and CPG signal; Design the CPG gait address generator.Secondly, we abstract quadruped robot dynamic walking into a invert pendulum model. Use the inverted pendulum models to analyze the longitudinal and transverse stability running conditions of a quadruped robot. Apply the CMAC theory to the stability control of quadruped robots. Results show that the CMAC theory can control the dynamic stability of the quadruped walking robot.Finally, we established a simulation platform of quadruped walking robot of dynamic walking. Verify the single foot trajectory generation algorithm, and according to the biological control mechanism of mammalian quadruped established the robot real-time motion control simulation system. The result of the simulation of quadruped robot dynamic trot gait show that CMAC neural network control method is reasonable and effective, can realize the quadruped robot stable dynamic walking.