Equipment System for Developing Natural Control Interface of Next Generation Affordable Prosthetic Hands

全世界有300万人失去了一只手臂，这严重影响了他们的日常生活。义肢在改善截肢者的生活方面有很大的潜力。目前可用的主动假肢可以根据其控制类型进行大致分类。非侵入式控制系统功能有限，界面缺乏直观控制，因为它们没有连接到人类的神经系统。假体的控制是通过使用模式识别技术解读残肢皮肤记录的肌电信号来调节的。这需要用户小心地运用不同的肌肉信号模式来执行不同的手势。由于检测运动意图的准确度较低，目前由非侵入性系统控制的假肢不能准确地执行期望的任务，导致肢体排斥。相反，可以通过手术将传感器连接到截肢者残肢的神经上来更好地控制假肢（例如，定向肌肉神经移植），但这是昂贵且有风险的（例如，手术并发症和感染）。因此，迫切需要提高非侵入性控制系统的鲁棒性，超越目前使用表面肌肉信号（即肌电信号）的限制。眼动追踪是一种非侵入性技术，在认知和教育研究中得到了广泛的应用。例如，检测眼睛对目标的注视有助于提取有关目标相对于用户的方向和距离的信息，并且眼睛的知识可以与识别和指导运动紧密相关。我们建议在假手的非侵入控制算法中加入眼动追踪数据，以提高对用户运动意图的理解准确性。具体来说，我们计划开发一种眼球追踪和肌电图系统，用于非侵入性手假体控制。我们的系统将包括：1)一个眼动仪，用于检测用户对环境中目标的视觉焦点；2)一个表面肌电图（sEMG）和力肌图（FMG）信号采集系统，用于检测残肢皮肤上的肌肉活动和模式，以及3)一个机器人手，用于触觉反馈和显示运动，该运动由我们的计算机算法根据从眼睛和肌肉收集的数据控制。这只机械手有内置的触觉传感器和振动致动器，用于触觉反馈。这三个组件将无缝集成到眼手跟踪和反馈系统中，并将作为我们研究平台的核心部分，我们可以在那里全面实时测试我们的假肢控制算法。我们的新技术发展对假肢和辅助设备控制以及广泛的人机交互具有重要意义。机器人领域的进步具有巨大的潜力，可以产生高影响的结果，并为加拿大社会做出巨大贡献。

Three million people around the world live with an arm lost, significantly impacting their daily life. Prosthetics have great potential to improve the lives of amputees. Currently available active prosthetics can be broadly classified based on their type of control. Non-invasive control systems have limited functions and the interface lacks intuitive control, as they are not connected to the human neural system. The control of the prosthesis is mediated by interpreting myoelectrical signals recorded from the skin of the residual limb using pattern recognition techniques. This requires the user to carefully exert distinct muscle signal patterns to perform different gestures. With a low accuracy in detecting movement intention, current prostheses controlled by non-invasive systems are not able to perform accurately desired tasks leading to limb rejection. Conversely, better control of a prosthesis can be achieved through surgery to connect sensors to the nerve in the amputee's residual limb (e.g., Targeted Muscle Reinnervation) but is expensive and risky (e.g., surgical complications and infections). Therefore, there is an urgent need to improve the robustness of non-invasive control systems beyond the current limitations of using surface muscle signals (i.e., myoelectric signals). Eye-tracking is a non-invasive technology that has been used extensively in cognitive and education studies. For instance, detecting the fixation of eyes on a target helps to extract information about the orientation and distance of the target relative to the user, and knowledge of the eyes can be tightly associated with identifying and guiding movements. We propose adding eye-tracking data to the non-invasive control algorithm of a prosthetic hand toward enhancing the accuracy in interpreting users' movement intention. Specifically, we plan to develop an eye-tracking and electromyogram system for non-invasive hand prosthesis control. Our system will include: 1) an eye-tracker for detecting users' visual focus on the target in the environment; 2) a surface Electromyography (sEMG) and Force Myography (FMG) signal acquisition system for detecting muscle activity and patterns from the skin on the residual arm, and 3) a robotic hand for haptic feedback and displaying movements controlled by our computer algorithm based on data collected from the eye and muscle. This robotic hand has built-in tactile sensors and vibration actuators for haptic feedback. These three components will be seamlessly integrated into an eye-hand tracking and feedback system, and will act as the core part of our research platform, where we can test our prosthesis control algorithms comprehensively in real-time. Our novel technological developments have important implications in prosthesis and assistive devices control, and a wide range of human-robot interactions. Advancements in the field of robotics have great potential to yield high impact results and make great contributions to Canadian society.