基于可穿戴式多模态人机接口的机械臂运动控制方法

A Motion Control Method for Robotic Arm Based on a Wearable Hybrid Human-Machine Interface

  • 摘要: 现有的人机接口系统存在指令较少、操作困难、任务能力受限等问题,无法有效拓展到机械臂的多维运动控制。本文提出了一种基于可穿戴式多模态人机接口的机械臂运动控制方法。该方法结合用户的眼电、头部姿态和语音等多模态信号,将其转换成控制指令,从而实现对机械臂在任意角度下的2维和3维连续运动控制。由10名受试者完成了指令输出、2维目标跟踪、字母书写和3维物体抓取等测试。结果显示,系统利用眨眼动作生成指令的平均准确率为96.67%,平均响应时间为1.51 s,平均信息传输率为142.53 bit/min,平均误报率为0.05次/分钟。此外,系统在2维平面沿2条不同路线跟踪目标的均方根偏差分别为0.12和0.14(归一化),抓取3维物体时的平均轨迹效率为92.65%,系统的控制效果与手动控制效果相当。实验结果验证了利用该多模态人机接口实现机械臂高效运动控制的可行性以及它在上肢运动功能辅助方面的应用潜力。

     

    Abstract: Existing HMI (human-machine interface) systems suffer from issues such as limited commands, complex operation, and restricted task capabilities, preventing effective expansion into multi-dimensional motion control for robotic arms. This paper introduces a method for controlling robotic arm movements based on a wearable hybrid HMI. This method combines various signals, including electrooculography (EOG), head posture, and speech from the user, transforming them into control commands, thereby enabling continuous two-dimensional (2D) and three-dimensional (3D) motion control of the robotic arm at any angle. 10 participants complete tests involving command output, 2D target tracking, alphabetic writing, and 3D object grasping. The results indicate that the blink-generated commands of the proposed system have an average accuracy of 96.67%, an average response time of 1.51 s, an average information transfer rate (ITR) of 142.53 bit/min, and an average false positive rate (FPR) of 0.05 event/min. Additionally, the root mean square deviations of target tracking along 2 different routes on a 2D plane are 0.12 and 0.14 (normalized), while the average trajectory efficiency of 3D object grasping is 92.65%. The control performance of the system is comparable to manual control. The experimental results verify the feasibility of using a hybrid HMI for achieving efficient motion control of robotic arms and its potential application in assisting upper-limb mobility functions.

     

/

返回文章
返回