基于刚度阻尼特征的神经网络自适应阻抗控制*

doi:10.16731/j.cnki.1671-3133.2024.12.001

现代制造工程 ›› 2024, Vol. 531 ›› Issue (12): 1-8.doi: 10.16731/j.cnki.1671-3133.2024.12.001

• 试验研究 • 下一篇

基于刚度阻尼特征的神经网络自适应阻抗控制^*

党选举^1,2, 黄伟健^1,2

1 桂林电子科技大学电子工程与自动化学院,桂林 541004;
2 广西智能综合自动化高校重点实验室,桂林 541004

收稿日期:2024-04-23 出版日期:2024-12-18 发布日期:2024-12-24
作者简介:党选举,博士,教授,主要研究方向为神经网络建模与控制、机器人智能控制。黄伟健,硕士研究生,主要研究方向为机器人力柔顺控制。E-mail:xjd69@163.com
基金资助:
^*国家自然科学基金项目(62263004);广西重点研发计划项目(桂科AB23075102)

Neural network adaptive impedance control based on stiffness damping characteristics

DANG Xuanju^1,2, HUANG Weijian^1,2

1 School of Electronic Engineering and Automation,Guilin University of Electronic Technology, Guilin 541004,China;
2 Key Laboratory of Guangxi College Intelligent Comprehensive Automation, Guilin 541004,China

Received:2024-04-23 Online:2024-12-18 Published:2024-12-24

摘要/Abstract

摘要： 针对打磨机器人阻抗控制的力跟踪性能受环境刚度未知和环境位置变化影响的问题,提出了一种基于刚度阻尼特征的神经网络自适应阻抗控制方法。由于环境参数未知导致参考轨迹不易确定,构造了一个自适应PI控制律,进行参考轨迹补偿,减小力跟踪的稳态误差;为了提高力跟踪控制的动态性能,根据力误差对刚度系数及阻尼系数的统一调节规律——刚度阻尼特征,并结合力误差具有时变、非线性的特点,设计了一个描述力误差与刚度阻尼特征关系的激活函数,构建自适应阻抗参数神经网络模型,其输出为刚度系数和阻尼系数,通过基于参考轨迹补偿与自适应阻抗参数神经网络模型融合的阻抗控制,保证力跟踪控制的柔顺性。仿真结果表明,相比于传统阻抗控制和参考轨迹PI补偿的阻抗控制,所提出的自适应阻抗控制方法具有更好的力跟踪效果。

关键词: 力柔顺控制, 未知环境, 参考轨迹补偿, 变阻抗参数模型, 神经网络

Abstract: A neural network adaptive impedance control method based on stiffness damping characteristics was proposed to address the problem that the force tracking performance of impedance control for sanding robots was affected by the unknown environmental stiffness and the change of environmental position.Since the reference trajectory is not easy to be determined due to the unknown environmental parameters,an adaptive PI control law was constructed to compensate the reference trajectory and reduce the steady-state error of force tracking; in order to improve the dynamic performance of the force tracking control,according to the uniform regulation law of the force error on the stiffness coefficient and damping coefficient—stiffness damping characteristics,and combined with that the force error has the characteristics of time-varying and non-linear,an activation function describing the relationship between force error and stiffness damping characteristics was designed,and an adaptive impedance parameter neural network model was constructed,whose outputs were stiffness coefficient and damping coefficient,to ensure the suppleness of force tracking control through the impedance control based on the fusion of the reference trajectory compensation and an adaptive impedance parameter neural network model. The simulation results show that the proposed adaptive impedance control method has better force tracking effect than the traditional impedance control and the impedance control with reference trajectory PI compensation.

Key words: force-flexing control, unknown environment, reference trajectory compensation, variable impedance parameter model, neural network

中图分类号:

TH165
TP242

党选举, 黄伟健. 基于刚度阻尼特征的神经网络自适应阻抗控制^*[J]. 现代制造工程, 2024, 531(12): 1-8.

DANG Xuanju, HUANG Weijian. Neural network adaptive impedance control based on stiffness damping characteristics[J]. Modern Manufacturing Engineering, 2024, 531(12): 1-8.

参考文献

[1] 李广,孔飞飞,王逍遥,等.基于多视觉传感器融合的装配过程质量移动智能检测技术研究[J].现代制造工程,2023(4):116-123.
[2] 徐天一,樊留群,王楠,等.基于激光追踪仪的机器人标定及误差分析研究[J].制造业自动化,2020,42(4):61-64.
[3] 李振,赵欢,王辉,等.机器人磨抛加工接触稳态自适应力跟踪研究[J].机械工程学报,2022,58(9):200-209.
[4] ROVEDA L,SHAHID A A,IANNACCI N,et al. Sensorless optimal interaction control exploiting environment stiffness estimation[J].IEEE Transactions on Control Systems Technology,2021,30(1):218-233.
[5] WU R,BILLARD A.Learning from demonstration and interactive control of variable-impedance to cut soft tissues[J].IEEE/ASME Transactions on Mechatronics,2021,27(5):2740-2751.
[6] LIU X,GE S S,ZHAO F,et al.Optimized interaction control for robot manipulator interacting with flexible environment[J].IEEE/ASME Transactions on Mechatronics,2020,26(6):2888-2898.
[7] 曾令城,李明富,杨真真,等.基于先验速度修正的工业机器人曲面跟踪柔顺控制[J].机械工程学报,2022,58(1):41-51.
[8] PAN Y,LI Z,SHI T,et al.Composite learning variable impedance robot control with stability and passivity guarantees[J].IEEE Robotics and Automation Letters,2023,9(1):119-126.
[9] ANAND A S,KAUSHIK R,GRAVDAHL J T,et al.Data-Efficient Reinforcement Learning for Variable Impedance Control[J].IEEE Access,2024,12(1):15631-15641.
[10] 郭万金,赵伍端,利乾辉,等.基于集成概率模型的变阻抗机器人打磨力控制[J].浙江大学学报(工学版),2023,57(12):2356-2366,2374.
[11] SUN T,PENG L,CHENG L,et al.Stability-guaranteed variable impedance control of robots based on approximate dynamic inversion[J].IEEE Transactions on Systems,Man,and Cybernetics:Systems,2019,51(7):4193-4200.
[12] 梁旭,王卫群,苏婷婷,等.下肢康复机器人的主动柔顺自适应交互控制[J].机器人,2021,43(5):547-556.
[13] 李冬武,张洁,汪俊亮,等.细纱接头机器人神经网络自适应力跟踪导纳控制[J].机械工程学报,2023,59(11):221-231.
[14] HUANG H,YANG C,CHEN C L P.Optimal robot-environment interaction under broad fuzzy neural adaptive control[J].IEEE Transactions on Cybernetics,2020,51(7):3824-3835.
[15] 郭万金,于苏扬,田玉祥,等.机器人打磨自适应变阻抗主动柔顺恒力控制[J].哈尔滨工业大学学报,2023,55 (12):54-65.
[16] 王晓华,王育合,张蕾,等.缝纫机器人对织物张力与位置的模糊阻抗控制[J].纺织学报,2021,42(11):173-178.
[17] LACHNER J,ALLMENDINGER F,STRAMIGIOLI S,et al.Shaping impedances to comply with constrained task dynamics[J].IEEE Transactions on Robotics,2022,38(5):2750-2767.
[18] LI G,YU J,CHEN X.Adaptive fuzzy neural network command filtered impedance control of constrained robotic manipulators with disturbance observer[J].IEEE Transactions on Neural Networks and Learning Systems,2021,34(8):5171-5180.
[19] 杨国成.基于力反馈的打磨机器人去毛刺作业系统研究[D].广州:广东工业大学,2022.

基于刚度阻尼特征的神经网络自适应阻抗控制^*

Neural network adaptive impedance control based on stiffness damping characteristics

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	马黎, 张迪. 基于QP-ZNN的冗余度机械臂容错控制^*[J]. 现代制造工程, 2025, 537(6): 73-83.
[2]	金奕扬, 李磊, 许家伟, 汪建华, 王国伟, 许润康. 基于RBF神经网络的4-PPPS并联机构位姿误差补偿^*[J]. 现代制造工程, 2025, 535(4): 140-150.
[3]	徐帅, 李艳武, 谢辉, 牛晓伟. 基于卷积金字塔网络的PPO算法求解作业车间调度问题^*[J]. 现代制造工程, 2025, 534(3): 19-30.
[4]	彭茂珲, 邹帅, 滕家皇, 黄福川. 改进北方苍鹰算法在立磨下摇臂结构优化的应用研究^*[J]. 现代制造工程, 2025, 533(2): 121-129.
[5]	李良辉, 李乐, 王茜, 张喜明. 基于DBO-BP的永磁同步电机损耗预测模型^*[J]. 现代制造工程, 2025, 533(2): 130-137.
[6]	胡鹰, 原嘉辰, 吕畅. 一种辊式矫直智能优化工艺预测模型的研究与应用^*[J]. 现代制造工程, 2024, 527(8): 109-117.
[7]	李泽稷, 周学良, 孙培禄. 融合残差块与Swin-Transformer机制的刀具磨损监测方法^*[J]. 现代制造工程, 2024, 527(8): 126-135.
[8]	张苗苗, 侯荣国, 吕哲, 王龙庆, 石广行, 王中庆. 基于GA-GRNN的AWJ强化3D打印AlSi10Mg表面性能实验研究^*[J]. 现代制造工程, 2024, 526(7): 35-41.
[9]	杨怀磊, 姚云磊. 基于故障观测器的机器人多关节手臂最优容错控制^*[J]. 现代制造工程, 2024, 526(7): 61-68.
[10]	邓旭, 张爱梅. 液压门式起重机受力及稳定性监测系统研究与应用[J]. 现代制造工程, 2024, 526(7): 135-143.
[11]	贾英霞, 王东辉. 基于自适应神经网络的工业机器人双臂协同鲁棒控制^*[J]. 现代制造工程, 2024, 525(6): 61-68.
[12]	聂鑫鹏, 袁逸萍, 祁雷, 朱广贺. 压延车间能源信息采集关键技术研究与系统实现[J]. 现代制造工程, 2024, 525(6): 75-81.
[13]	成元彬, 袁文平, 张涛, 刘志峰. 基于BP神经网络的谐波减速器柔轮疲劳寿命预测研究^*[J]. 现代制造工程, 2024, 523(4): 140-145.
[14]	张若语, 胡友民, 吴波, 杨晔, 秦峻峰. 基于BA-PNN算法与数字孪生的车间扰动判定方法[J]. 现代制造工程, 2024, 522(3): 15-22.
[15]	孙建民, 杨世虎, 赵磊, 姚德臣. 基于动态惯性权重的电子节气门改进PSO-BP优化控制^*[J]. 现代制造工程, 2024, 521(2): 45-52.