基于改进门控循环神经网络的机械臂逆运动学解  

Inverse Kinematic Solution of Robotic Manipulator Based on Improved Gated Recurrent Neural Network

作　　者：陈富安 侯宇杰 李苏阳 CHEN Fu’an;HOU Yujie;LI Suyang(College of Electrical Engineering,He’nan University of Technology,He’nan Zhengzhou 450001,China)

机构地区：[1]河南工业大学电气工程学院,河南郑州450001

出　　处：《机械设计与制造》2025年第1期365-368,373,共5页Machinery Design & Manufacture

基　　金：河南省科技攻关项目(182102210088)。

摘　　要：针对传统神经网络和智能算法在机械臂逆运动学求解过程中求解精度低和泛化能力弱的问题,提出一种利用注意力机制改进门控循环(Gated Recurrent Unit,GRU)神经网络的逆运动学求解模型。该模型利用GRU神经网络对机械臂逆向运动过程中时序性的分析,学习各个关节运动前后变化的约束关系。然后通过注意力机制突出重要节点数据变化对关节角的影响,提高求解精度,增强模型泛化能力。本仿真以六自由度串联机械臂为研究对象,对其运动学进行建模分析,将机械臂的逆运动学求解问题转换为误差最小化的研究。仿真结果表明,利用该神经网络模型求解逆运动学相比于PSO-BP和LSTM网络,模型的求解精度得到了明显提高。For the problem of low accuracy and weak generalization ability in traditional neural network and intelligent algo⁃rithm,an inverse kinematic solution model for improved neural networks of gating cycle(Gated Recurrent Unit,GRU)using the attention mechanism is proposed.The model uses the GRU neural network analyzing the timing of the mechanical arm to learn the constraint of the forward and backward changes in each joint.The addition of attention mechanism highlights the impact of important node data changes on changes in joint angle to improve the generalization ability of the model.In this paper,the sixaxis robot arm will model the forward kinematics principle and transform the inverse kinematic solution problem of the robot arm into the study of error minimization.Simulation results show that using the neural network to solve inverse kinematics can signifi⁃cantly improve the model kinematics compared to PSO-BP and LSTM networks.

关 键 词：逆运动学 注意力机制 GRU神经网络 求解精度 

分 类 号：TH16[机械工程—机械制造及自动化] TP242[自动化与计算机技术—检测技术与自动化装置]