Pattern Recognition Based Myoelectric Control for Emerging Human-Machine Interfaces

技术的可得性和实用性的迅速变化使社会处于不断适应的状态。现在，许多人意识到久坐不动的生活方式以及电脑和移动设备的普遍使用带来的危险，他们都在争先恐后地寻找替代品。随着头戴式虚拟现实和增强现实设备的出现，人们对可以在旅途中进行交互的免提交互方法的需求不断增长。几十年来，肌电控制，利用肌肉收缩过程中产生的电信号来解码使用者的意图，已被用于控制截肢者的假肢。最近，技术、机器学习和用户培训的进步彻底改变了肌电控制的实现方式，使对多个自由度的控制更加直观。然而，相比之下，肌电控制作为潜在的人机界面用于其他消费产品的探索相对较少。那些追求这一方向的人大多直接应用了假肢研究机构的知识，限制了其有效性。因此，对于消费者使用肌电控制的潜力、稳健性和直观性，存在许多悬而未决的问题。

Rapid changes in the availability and utility of technology have left society in a state of constant adaptation. Now realizing the dangers of sedentary lifestyles and pervasive computer and mobile device use, many are scrambling to find alternatives. With the advent of emerging heads-up virtual and augmented reality devices, there is growing demand for hands-free interaction approaches that could enable interaction on the go. For decades, myoelectric control, the use of electrical signals generated during muscle contractions to decode user intent, has been used to control prostheses for amputees. Recently, advances in technology, machine learning, and user training have revolutionized the way myoelectric control is achieved, enabling more intuitive control of multiple degrees of freedom. In comparison, however, myoelectric control has been relatively unexplored as a potential human-machine interface for use with other consumer products. Those who have pursued this direction have mostly directly applied the knowledge from the body of prosthetics research, limiting its effectiveness. Consequently, there are many outstanding questions about the potential, robustness and intuitiveness of myoelectric control for consumer use. In this work, we will develop new human-machine interaction models and devices specifically designed for use with emerging consumer hands-free technologies. As the Innovation Research Chair at the Institute of Biomedical Engineering, a world-renowned leader and influencer in the field of myoelectric control, we will leverage lessons learned in developing the state-of-the-art in myoelectric control for prosthetics, while designing for this very different use-case. We will determine appropriate gesture sets that balance intuitiveness for the application with the information content and robustness of the control scheme and improve the training and learning process by considering the user as part of the control system. Finally, we will explore alternative sensors and the role of sensor fusion in developing versatile and robust hands-free human-machine interfaces. This work will present engaging multidisciplinary training opportunities for students and develop highly sought-after skillsets in problem solving, critical thinking, signal processing, machine learning, and human machine interaction. The results of this program will produce foundational research that will broaden the applications for myoelectric control and inform the design of novel technologies. The development of intuitive, hands-free human machine interfaces will reduce the restrictions and risks associated with current work environments.