Man-machine interfacing based on ultrasound wearable technology for controlling upper limb prostheses

当前用于上肢假肢的最先进的控制方法是基于用非侵入性电极（表面EMG）检测的肌肉电活动。然而，表面EMG具有差的空间分辨率，受皮肤-电极界面的影响，具有限于几毫米至1厘米深度的检测体积（因此不能进入深层肌肉），并且由于疲劳和其他生理因素而改变其特性。这些局限性在主动假肢的传统肌控制策略中提出了重要问题，与大约50%的放弃率相关。本项目旨在研究一种替代的、临床上可行的和非侵入性的方法-肌肉超声传感和成像-用于检测肌肉骨骼结构的运动，并在患者和假肢/手之间建立神经接口。为此，我们将开发一种可穿戴超声系统来检测和表征肌肉结构的活动和弹性状态，并开发机器学习方法来将超声信号映射为仿生肢体的命令。超声技术可以以可扩展的空间分辨率（数十微米至~ 1 mm）深入软组织（几厘米至15厘米），并在检测细微的局部组织运动方面表现出出色的灵敏度。超快超声技术的最新进展使得微米级的深部组织运动检测能够以<1毫秒的时间分辨率进行检测，因此超声可以通过图像监测整个肌肉横截面来提供肌肉运动的直接测量，具有非常高的时间分辨率，并且不受影响EMG记录的混杂因素的影响。此外，超声系统可以小型化，因此可以作为可穿戴设备安装。该项目将首次开发用于可穿戴人机接口的超声和信号处理技术，并将这些发展转化为临床可用性测试。研究小组和博士目标正在进行的工作。由Dario Farina教授领导的伦敦帝国理工学院生物工程系的神经力学和康复技术小组专注于研究运动神经控制的电生理技术、生物电极和生物信号处理、神经康复、主动假肢、人机界面和运动神经元体内记录。该研究小组通过开发概念和技术，填补了人类运动的神经和生物力学研究之间的差距，对肌肉电生理学和人机界面做出了重大贡献。成像和传感超声实验室（ULIS）由Mengxing Tang教授领导，主要致力于为广泛的生物医学应用开发新的超声成像和传感技术。

Current state-of-the-art control methods for upper limb prostheses are based on muscle electrical activity detected with non-invasive electrodes (surface EMG). However, surface EMG has poor spatial resolution, is influenced by the skin-electrode interface, has a detection volume limited to few millimetres up to 1 centimetre depth (therefore not accessing deep muscles), and varies its characteristics as a consequence of fatigue and other physiological factors. These limitations pose important issues in conventional myo-control strategies for active prostheses, associated to an abandonment rate of approximately 50%. This project aims to investigate the usage of an alternative, clinically-viable and non-invasive approach -muscle ultrasound sensing and imaging --for detecting the motion of musculoskeletal structure sand establishing a neural interface between patients and prosthetic arms/hands. For this purpose, we will develop a wearable ultrasound system to detect and characterize the activity and elastic status of muscle structures and machine learning methods to map the ultrasound signals into commands for bionic limbs. Ultrasound techniques can penetrate deep in soft tissue (a few cm to15cm) with scalable spatial resolution (tens of microns to ~1mm), and have shown excellent sensitivity in detecting subtle local tissue motion. Recent advances in ultrafast ultrasound have enabled the detection of deep tissue motion at micrometre level to be detected at a temporal resolution <1 millisecond.Hence ultrasound can provide a direct measure of muscle movement by image monitoring the entire muscle cross-sectional are a with very high temporal resolution, and is not influenced by the confounding factors affecting EMG recordings. Moreover, ultrasound systems can be miniaturized and therefore mounted as wearable devices. The project will develop for the first time ultrasound and signal processing technology for wearable man-machine interfacing and will translate these developments into clinical usability tests.Ongoing work in the research group and the PhD aims. The Neuromechanics and Rehabilitation Technology group at the Department of Bioengineering of Imperial College London, led by Prof. Dario Farina, focuses on electrophysiology techniques for the study of neural control of movement, bioelectrodes and biosignal processing, neurorehabilitation, active prostheses, human-machine interfaces, and motor neuron recordings in vivo. There search of the group has made major contributions to muscle electrophysiology and human-machine interfacing by developing concepts and techniques to fill the gap between the neural and biomechanical investigation of human movement. The Ultrasound Laboratory for Imaging and Sensing (ULIS), led by Professor Mengxing Tang, mainly focuses on developing new ultrasound imaging and sensing techniques for a wide range of biomedical applications.