School of Electrical Engineering, Henan University of Technology, Zhengzhou 450001, China
Show less
文章历史+
Received
Published
Revised
2025-08-02
2026-07-08
2025-12-09
Online Date
2026-08-06
PDF (18875K)
摘要
针对传统机械臂在复杂场景下交互效率低、控制方式不够直观的问题,设计了一种基于视觉感知的多维手势识别机械臂交互控制系统,该系统可实现动静态手势的3种交互控制功能:手部二维关节姿态的静态手势交互、手部末端三维运动的动态手势交互,以及手部运动轨迹的跟随控制交互,从而达成机械臂的直观高效操控。其中,静态手势关节交互采用改进的YOLO(you only look once)结合MediaPipe框架,通过检测手部关键点实现关节精准控制,识别准确率达93.9%;动态手势末端运动交互借助3DResNet与双向长短时记忆网络(bidirectional long short-term memory,Bi-LSTM)提取动态手势的时空特征,识别准确率达93.87%,可驱动机械臂末端按预定义指令运动;手部运动轨迹跟随控制交互利用深度相机捕获三维坐标,结合卡尔曼滤波和滑模控制器实现平滑轨迹跟踪,误差控制在4.5 mm以内。通过在UR5和Dobot机械臂上开展仿真与实物实验,验证了这3种交互功能的有效性与准确性,为机械臂人机交互提供了高效、自然且跨平台的解决方案。
Abstract
To address the issues of low interaction efficiency and non-intuitive control methods of traditional robotic arms in complex scenarios, a robotic arm interaction control system based on multi-dimensional gesture recognition under visual perception was designed. Three types of interaction control functions were accomplished, including static gesture interaction for the two-dimensional joint posture of the hand, dynamic gesture interaction for the three-dimensional movement of the hand’s end, and following control interaction for the hand’s movement trajectory. In this way, intuitive and efficient robot manipulation was enabled.The static gesture joint interaction was realized by combining an improved YOLO (you only look once) with the MediaPipe framework to achieve precise joint control through the detection of hand key points, with an identification accuracy rate of 93.9%. The dynamic gesture end movement interaction was implemented using 3DResNet and Bi-LSTM(bidirectional long short-term memory) networks to extract the spatiotemporal features of dynamic gestures, achieving an identification accuracy rate of 93.87%, and was used to drive the robotic arm’s end to move according to predefined instructions. The hand movement trajectory following control interaction was achieved by capturing three-dimensional coordinates with a depth camera and combining Kalman filtering and sliding mode controllers to ensure smooth trajectory tracking, with an error controlled within 4.5 mm.Through simulation and physical experiments conducted on UR5 and Dobot robotic arms, the effectiveness and accuracy of these three interaction functions are verified, providing an efficient, natural, and cross-platform solution for human-robot interaction.
ZHANG Jia-jian, WU Xiang, LI Bo, WANG Jun-chao, BAO Dong-hao.
Multi-dimensional Dual-state Gesture Interaction System for Robotic Arms Based on Visual Perception[J].
Science Technology and Engineering, 2026, 26(19): 8005-8018 DOI:10.12404/j.issn.1671-1815.2506127
本系统是一种通过手势识别实现机器人控制的人机交互系统,旨在提升人机交互的自然性与智能化水平。系统支持3种交互控制模式:静态手势关节交互、动态手势末端运动交互以及手部运动跟随控制交互,以适应不同的应用场景和交互需求。在静态手势关节交互模式中,系统基于改进的YOLO网络并结合MediaPipe框架实现对特定静态手势的识别,进而映射为对应的机器人关节角控制信号,实现关节级别的精细交互。在动态手势末端运动交互模式中,系统引入3DResNet提取视频序列中的时空特征,并结合双向长短期记忆网络(bidirectional long short-term memory,Bi-LSTM)对时间序列信息进行建模,提升对动态手势的识别准确率和鲁棒性。识别结果将被映射为预定义的机器人任务指令。在手部运动跟随交互模式中,系统利用深度相机(如RealSense)实时采集手部的三维空间轨迹信息,通过卡尔曼滤波器对采集的轨迹进行平滑与去噪处理,从而生成稳定的运动路径。随后,系统将手部轨迹映射至机器人末端执行器的空间轨迹,实现对手部运动的同步跟随控制,增强交互的直观性与灵活性。
LiuShaoshan, GanYiming, HanYinhe. Research on the development of robot computing[J]. Bulletin of the Chinese Academy of Sciences, 2024, 39(11): 1956-1965.
WangChang, ZhangYan, ZhaoChen, et al. The influence characteristics of a new full-touchscreen human-machine interaction mode on driving risk[J]. Science Technology and Engineering, 2023, 23(22): 9673-9679.
YaoXin. Research on key technologies of multimodal human-computer interaction system based on speech recognition[J]. Automation and Instrumentation, 2023 (11): 222-225.
LongZhang, LuChengya, LiGuopeng, et al. Design of a mobile robot gesture control system based on visual gesture recognition[J]. Mechanical Science and Technology, 2022, 41(8): 1247-1253.
[9]
PavithraR, ShreejaP, SirishaM V K, et al. Gesture control of robotic arm for hazardous environment[J]. International Journal of Engineering Trends and Technology, 2018, 57(1): 1822.
[10]
ZhangW, ChengH, ZhaoL, et al. A gesturebased teleoperation system for compliant robot motion[J]. Applied Sciences, 2019, 9(24): 5290.
[11]
PeralM, SanfeliuA, GarrellA. Efficient hand gesture recognition for human-robot interaction[J]. IEEE Robotics and Automation Letters, 2022, 7(4): 10272-10279.
[12]
HeD, YangY, WuR. Design of human-computer interaction gesture tracking model based on improved PSO and KCF algorithms[J]. IEEE Access, 2024, 12: 39913-39925.
[13]
XuJ, WangH, ZhangJ, et al. Robust hand gesture recognition based on RGB-D data for natural human-computer interaction[J]. IEEE Access, 2022, 10: 54549-54562.
[14]
JiaL, ZhouX, XueC. Non-trajectory-based gesture recognition in human-computer interaction based on hand skeleton data[J]. Multimedia Tools and Applications, 2022, 81(15): 20509-20539.
GuoJian, ZhangShu, WangShuxuan, et al. Gesture recognition algorithm for millimeter-wave radar based on Vision Transformer[J]. Modern Radar, 2023, 45(12): 79-86.
[17]
ZhangR, JiangC, WuS, et al. Wi-Fi sensing for joint gesture recognition and human identification from few samples in human-computer interaction[J]. IEEE Journal on Selected Areas in Communications, 2022, 40(7): 2193-2205.
[18]
WuX, MaY, ZhangS, et al. Yo3RL-Net: a fusion of two-phase end-to-end deep net framework for hand detection and gesture recognition[J]. Alexandria Engineering Journal, 2025, 12: 177-189.
ZhengHan, TianMeng, ZhaoYanfeng, et al. Hand pose estimation method based on improved Faster R-CNN[J]. Science Technology and Engineering, 2023, 23(3): 1160-1167.
[21]
ZhangJ, MengY, YuX, et al. MBAB-YOLO: a modified lightweight architecture for real-time small target detection[J]. IEEE Access, 2023, 11: 78384-78401.
QiaoJinghui, LiHongda, ChenYuxi, et al. Mechanism and data-driven soft measurement model of soft hand bending angle[J]. Journal of Electronic Measurement and Instrumentation, 2023, 37(6): 147-158.
[24]
MittalA, ZissermanA, TorrP H S. Hand detection using multiple proposals[C]// Proceedings of the British Machine Vision Conference. Dundee: BMVA Press, 2011: 75.
[25]
PandaA K, ManikandanM S, RamkumarB. Bidirectional LSTM based sequence classification approach for human presence detection using channel state information[C]// 2023 International Conference on Microwave, Optical, and Communication Engineering (ICMOCE). New York: IEEE, 2023: 1-6.
[26]
MaterzynskaJ, BergerG, BaxI, et al. The jester dataset: a large-scale video dataset of human gestures[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops. Seoul: IEEE, 2019: 2874-2882.
[27]
ElangovanT, Arockia Xavier AnnieR, SundaresanK, et al. Hand gesture recognition for sign languages using 3DCNN for efficient detection[C]// Computer Methods, Imaging and Visualization in Biomechanics and Biomedical Engineering II. Berlin: Springer International Publishing, 2022: 215-233.
[28]
ChandrabanshiV, DomnicS. A novel framework using 3D-CNN and BiLSTM model with dynamic learning rate scheduler for visual speech recognition[J]. Signal, Image and Video Processing, 2024, 18(6): 5433-5448.
[29]
PrasanthN, ShrivastavaK, SharmaA, et al. Gesture-based mouse control system based on MPU6050 and Kalman filter technique[J]. International Journal of Intelligent Systems Technologies and Applications, 2023, 21(1): 56-71.
QiangLili. Research on intelligent detection and tracking of movement trajectory of automatic object finder robot in library[J]. Mechanical Design and Manufacturing Engineering, 2023, 52(5): 43-48.
ZhaoGongmei, GuXianyu. Motion trajectory reconstruction method based on infrared sensor[J]. Journal of Lasers, 2021, 42(9): 155-160.
[34]
MedinaL, GuerraG, HerreraM, et al. Trajectory tracking for non-holonomic mobile robots: a comparison of sliding mode control approaches[J]. Results in Engineering, 2024, 22: 102105.