General calculation method for inverse kinematics of serial robots

Abstract

The invention discloses a general calculation method for inverse kinematics of serial robots. The general calculation method comprises the following steps of classifying and simplifying a plurality of types of robot moving joints to establish a uniform robot moving joint kinematic model; performing two-dimensional processing on robot working configuration according to a planar processing method, classifying and dividing formed planar configurations on the basis and respectively establishing automatic planar configuration working space calculation methods; and performing configuration plane matching on any robot configuration to solve the inverse kinematics of the serial robots. The method overcomes the defects of the limitation and specificity for solving the robot configuration by the traditional analytic method, and solves the problems of non-instantaneity and low accuracy of a general iterative method, can quickly and accurately solve the inverse kinematics of the robots, and meets the requirement for calculating the general robot kinematics and the actual requirement on robot motion control.

Description

technical field [0001] The invention relates to a general solution method for the inverse kinematics of a series robot, in particular a solution method that integrates the general modeling method of robot joints, configuration plane division and configuration plane matching. Background technique [0002] The problem of robot kinematics is the basic problem of robot motion control, and the general and analytical form of inverse kinematics of serial robots has always been a difficult problem in robotics research, and it is also a hot issue in the field of robotics research. [0003] There are two main directions for solving robot kinematics problems: analytical methods and numerical methods. The D-H modeling analysis method is an earlier method in the analysis method, but due to the large amount of calculation and complexity, and it depends on the configuration of the robot, this method is usually aimed at the fixed robot structure with few degrees of freedom. The shape is ca...

AI Technical Summary

Problems solved by technology

According to the screw theory, the modeling of robot joints does not depend on the configuration of the robot, but although the formulas of screw and exponential product provide a unified form of forward kinematics that is independent of the type of joints and has a concise form, and in kinematics On the basis of the positive solution, a closed-form geometric method of inverse kinematics can be constructed, but the entire inverse kinematics problem needs to be decomposed into several solvable sub-problems, the solution process is complicated, and it is difficult to automatically solve kinematics, and this This solution method is also not universal, and is only applicable to certain forms of robot working configurations

The more commonly used method in the numerical method is to use the Newton-Raphson iterative method to perform iterative calculations based on the established forward motion model to obtain the inverse kinematics results of the robot, but it is difficult to guarantee the accuracy of the calculation time and results at the same time.

Genetic algorithms and neural network methods can optimize the solution, avoid local convergence, and obtain inverse kinematics results, but like iterative methods, there are constraints on calculation time and calculation accuracy

Images