期刊
出版社
SAGE PUBLICATIONS LTD
DOI: 10.1177/09544062231184791
关键词
Redundant manipulator; inverse kinematics; neural network algorithm; gravitational search algorithm
In this study, a cascade calculation strategy combining global mapping and local search is proposed to solve the inverse kinematics problems of the redundant hydraulic manipulator of the demolition robot. The inverse kinematics solution is mapped from the global space to the local space of the working point using a global neural network model, and then the inverse solution results are obtained through the local search of the gravitational search algorithm. An improved gravitational particle adaptive step search strategy is proposed to balance the exploration and development stage of the population and avoid local minima. Simulation and experimental results demonstrate that the global-to-local inverse solution algorithm can avoid multiple solution problems and ensure the stability and accuracy of the solution based on real-time inversion.
In solving the inverse kinematics of the redundant hydraulic manipulator of the demolition robot, there are problems of multiple solutions and the contradiction between the real-time and accuracy of the calculation. A series calculation strategy combining global mapping and local search was proposed. Firstly, the inverse kinematics solution was mapped from the global space to the local space of the working point by the global neural network model, and then the inverse solution results were obtained by the local search of the gravitational search algorithm. To avoid the GSA falling into the local minima and improve the search speed and accuracy, an improved gravitational particle adaptive step search strategy was proposed to balance the exploration and development stage of the population. The simulation and experimental results show that the global than local inverse solution algorithm can avoid the multi-solution problems and ensure the stability and accuracy of the solution on the basis of real-time inversion.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据