Position And Pose Error Analysis Of Multilegged Walking Robots

The multilegged walking robot has outstanding advantages such as abundant motion, adaptability, and high carrying capacity. Positioning accuracy is one of the most important working performances of robot, which affect the working quality of the multilegged walking robot directly. Based on the motion analysis of a quadruped robot, the posture error analysis and accuracy compensation of the robot is discussed in this dissertation.Based on analyzing the Mechanism of the multilegged walking robot, the main error source is found. The errors can be modeled equivalently to several major sources of error according to the kinematics analysis of robot. These factors boil down to the length errors of legs, angle errors of driven joints, and position errors of footholds.On the basis of forward kinematics analysis, the forward kinematics error model for the posture error analysis of the multilegged walking robot has been formulated by the differential method of robots. The process consists of solving the errors of redundant actuation joints, error of direction matrix and the error of position. The programs for error analysis are compiled in MATLAB environment, the correctness of error analysis model is verified through the examples of simulation.The effect of error sources on the posture error of robot is analyzed. Considering each error factor as the variables respectively, the change of posture error is studied. The error characteristic curves are obtained by giving the input parameters. It is showed that the errors of driven joints have the largest influence on robot's posture error, and that the pose of robot also has a significant influence on the posture errorA method of accuracy compensation of multilegged walking robots is then introduced in the paper. The angles of driven joints are chosen to be the compensation variables for they are input variables. The compensation quantities of driven joints are obtained by solving compensation model of multilegged walking robots. The instance simulation demonstrates the feasibility and effectiveness of the compensation method.