Simultaneous Control of Multiple Degrees of Freedom in Myoelectric Hand Prostheses (SimCon)

这个项目的目的是开发一种全新的、基于生物信息的控制方法，能够同时控制上肢假体中的多个关节。任何肢体的丧失，特别是手的丧失，都会深刻地影响个人的生活质量。先进的假手可以通过允许截肢者进行日常活动来提供一条功能康复的途径，并增加他们重返职业生涯和赚取正常生计的机会。关于假手使用情况的调查显示，20%的截肢者放弃了假手，一个关键原因是它不能提供足够的功能。因此，提高上肢假肢的性能是一项迫切的需求。Reinhold Reiter在1945年的一项专利申请中披露的开关1自由度控制范例仍然被广泛用于假肢控制。早在1967年，Finley就证明了开关控制不能给使用者提供足够的灵活性，并提出使用模式识别来通过处理肌肉的电活动来估计假体使用者的运动意图，即肌电或肌电信号。今天，在芬利的提议50年后的今天，尽管有出色的实验室演示，但在肌电假手中实现模式识别商业化是不可行的，因为1)截肢者很难为不同的运动类别产生不同的活动模式，2)模式识别性能经常由于电极移位和残肢的移动而恶化。在这个项目中，我们将首先探索手和前臂的肌肉可以在多大程度上学习产生新的协同收缩模式(又名肌肉协同)，因为自然的协同作用可能会被截肢破坏。从这项实验工作中获得的洞察力将为设计能够同时控制多个关节(自由度)运动的新算法提供依据。当用户与假肢互动时，这些算法可以自我调整，以提高性能。该项目将在临床前试验中达到高潮，在该试验中，四名截肢者用假肢测试原型控制算法。将所提出的范例的性能与传统的假体开-关控制方法进行比较。提出的研究项目将为多自由度的同时控制提供一种有效的方法。这为用户提供了比当前基于开关或模式识别的控制方法更大的灵活性。试验结果表明，采用这种方法，假肢在100ms内就能响应使用者的运动意图，比目前最先进的上肢假肢控制至少快三倍。这项研究的结果将为下一代具有“即插即用”和“佩戴者独立”特征的“即插即用”假体铺平道路。

The aim of this project is to develop a radically novel and biologically-informed control approach that enables simultaneous control of multiple joints in an upper-limb prosthesis.The loss of any limb, particularly the hand, affects an individual's quality of life profoundly. Advanced prosthetic hands can provide a route to functional rehabilitation by allowing the amputees to undertake their daily activities and improve their chances of returning to their careers and earning their regular livelihood. Surveys on the use of artificial hands reveal that 20% of the amputees abandon their prosthesis with a key reason being that it does not provide enough function. Therefore, performance enhancement of upper-limb prostheses is a pressing need.The on-off 1-degree of freedom control paradigm that Reinhold Reiter disclosed in a patent application in 1945 is still used widely for prosthesis control. As early as 1967, Finley showed that the on-off control does not offer enough flexibility to the user and proposed the use of pattern recognition to estimate prosthesis user's movement intention by processing electrical activity of muscles, known as the electromyogram or myoelectric signals. Today, 50 years after Finley's proposal and despite remarkable laboratory demonstrations, it has not been feasible to commercialise pattern recognition in a myoelectric prosthesis hand because 1) it is very difficult for the amputees to generate distinct activity patterns for different movement classes and 2) pattern recognition performance often deteriorates due to electrode displacement and movement of the residual limb.In this project, we will firstly explore the extent to which muscles in the hand and forearm can learn to generate novel co-contraction patterns (aka muscle synergies) because natural synergies may be disrupted by amputation. The insight gained from this experimental work will inform design of novel algorithms to enable simultaneous control of multiple joints (degrees of freedom) movements. These algorithms can self-tune, to improve performance, as the user interacts with the prosthesis. The project will culminate in a pre-clinical trial in which four amputee subjects test the prototyped control algorithm with a prosthesis. The performance of the proposed paradigm will be compared to that of the conventional prosthesis on-off control method. The proposed research project will produce an efficient approach for simultaneous control of multiple degrees of freedom. This offers the user much greater flexibility than current on-off or pattern recognition-based control approaches. Pilot results show that with this approach the prosthesis can respond to user's motion intention in only 100ms, which is at least three times faster than the state-of-the-art in upper-limb prostheses control. Results of this research will pave the way for future generations of "plug and play" prostheses with "ready-to-go" and "wearer-independent" features.