[1]姚学恒,邢海青,杨子赫,等.基于惯性测量和LSTM算法的四足机器人上坡足端姿态控制[J].工业仪表与自动化装置,2025,(02):76-81.[doi:10.19950/j.cnki.CN61-1121/TH.2025.02.014]
　YAO Xueheng,XING Haiqing,YANG Zihe,et al.Attitude control of quadruped robot uphill foot end based on inertial measurement and LSTM algorithm[J].Industrial Instrumentation & Automation,2025,(02):76-81.[doi:10.19950/j.cnki.CN61-1121/TH.2025.02.014]

点击复制

基于惯性测量和LSTM算法的四足机器人上坡足端姿态控制(PDF)

《工业仪表与自动化装置》[ISSN:1000-0682/CN:61-1121/TH]

卷:

期数:

2025年02期

页码:

76-81

栏目:

出版日期:

2025-04-15

文章信息/Info

Title:

Attitude control of quadruped robot uphill foot end based on inertial measurement and LSTM algorithm

文章编号:

1000-0682（2025）02-0076-06

作者:

姚学恒; 邢海青; 杨子赫; 等

（1.国网浙江省电力有限公司杭州市余杭区供电公司；2.杭州申昊科技股份有限公司；3.浙江省智能运维机器人重点实验室，浙江 杭州 310000）

Author(s):

YAO Xueheng;  XING Haiqing;  YANG Zihe;  et al

(1.State Grid Zhejiang Electric Power Company Hangzhou Yuhang Power Supply Company; 2.Hangzhou Shenao Technology Co., Ltd.; 3.Zhejiang Key Laboratory of Intelligent Operation and Maintenance Robot, Zhejiang Hangzhou 310000, China)

关键词:

惯性测量; LSTM算法; 四足机器人; 足端姿态; 模型预测控制; 坐标系统

Keywords:

inertial measurement;  LSTM algorithm;  quadruped robot foot posture;  model predictive control;  coordinate system

分类号:

TP242

DOI:

10.19950/j.cnki.CN61-1121/TH.2025.02.014

文献标志码:

A

摘要:

为避免四足机器人在上坡过程中产生倾覆或滑倒，研究基于惯性测量和LSTM算法的四足机器人上坡足端姿态控制方法。利用配置加速度计和陀螺仪的惯性测量单元采集四足机器人足端上坡运动时的加速度和角速度数据，经四元数法以及积分运算获取机器人上坡足端姿态数据，将其作为LSTM（长短期记忆网络）模型的输入，预测机器人期望运动速度与位置；将其与惯行测量单元获取的实时足端位置与速度共同输入模型预测控制（MPC）器中，通过计算四足机器人摆动腿的关节力矩，结合其运动学模型，不断获取机器人坐标系下摆动腿足端的位置，完成四足机器人在上坡过程中的足端姿态控制。实验结果显示：该方法应用下，可以保障四足机器人顺利通过坡底到坡面、坡面到坡顶的过渡阶段，不产生打滑现象，保障上坡过程稳定。

Abstract:

To avoid tipping or slipping of quadruped robots during uphill climbing, a foot end attitude control method for quadruped robots based on inertial measurement and LSTM algorithm is studied. Using an inertial measurement unit equipped with accelerometers and gyroscopes to collect acceleration and angular velocity data of a quadruped robot during uphill motion, the robot’s uphill foot posture data is obtained through quaternion method and integration operation, and used as input for the LSTM (Long Short Term Memory Network) model to predict the expected motion speed and position of the robot; It is input into the Model Predictive Control (MPC) system along with the real-time foot position and velocity obtained from the inertial measurement unit. By calculating the joint torque of the swinging leg of the quadruped robot and combining it with its kinematic model, the position of the swinging leg foot in the robot coordinate system is continuously obtained to complete the foot posture control of the quadruped robot during uphill climbing. The experimental results show that under the application of this method, the quadruped robot can smoothly pass through the transition stages from the bottom of the slope to the slope surface and from the slope surface to the top of the slope without slipping, ensuring the stability of the uphill process.

参考文献/References:

[1] 余捷,刘绪清,于超,等.基于模型预测控制的四足机器人面向复杂地形自适应调整算法与实现[J].机电工程技术,2024,53(05):122-128.
[2] 赵江波,龚思进,王军政.四足机器人步态参数优化及探索性行走策略[J].北京理工大学学报,2022,42(04):407-414.
[3] 李姗姗,王洪波,陈建宇,等.新型四足并联军用机器人步态控制算法及仿真[J].兵工学报,2023,44(03):895-909.
[4] 刘帅,赵慧,刘清宇.四足机器人软硬地面稳定过渡的腿部主动变刚度调节策略[J].工程科学学报,2022,44(03):420-429.
[5] 赵杰亮,邢晋,牛群.基于俯仰位姿调整的四足机器人爬坡运动策略[J].华中科技大学学报(自然科学版),2023,51(06):10-15+22.
[6] 梁启星,李彬,李志,等.基于模型预测控制的四足机器人斜坡自适应调整算法与实现[J].山东大学学报(工学版),2021,51(03):37-44+51.
[7] Biswal P , Mohanty P K .Modeling and Effective Foot Force Distribution for the Legs of a Quadruped Robot[J].Robotica, 2021,39(8):1504-1517.
[8] 宫赤坤,吴浔炜,袁立鹏.基于阻抗和虚拟模型的四足机器人控制方法[J].系统仿真学报,2022,34(10):2152-2161.
[9] 许鹏,邢伯阳,刘宇飞,等.基于扩张状态观测器和模型预测方法的四足机器人抗干扰复合控制[J].兵工学报,2023,44(S2):12-21.
[10] 王银浩,颉潭成,徐彦伟,等.四足机器人侧摆型trot步态的运动学分析及仿真研究[J].现代制造工程,2022,(05):24-29.
[11] 于建均,李晨,左国玉,等.基于LSTM神经网络的仿人机器人循环步态的生成方法[J].北京工业大学学报,2020,46(12):1335-1344.?

相似文献/References:

备注/Memo

备注/Memo:

收稿日期：2024-07-18第一作者：姚学恒（1984—），男，汉族，浙江安吉人，硕士，高级工程师，主要研究方向为电网自动化运维、设备智能运检等。

常用功能