基于柔性传感的肌肉形变检测及其在运动意图识别中的应用

The commercial intelligent prosthesis is mainly controlled by surface electromyography (sEMG) signal. In practical application, it is easy to cause poor quality of sEMG signal due to sweating on skin surface, change of electrode position, environment electromagnetic interference and so on, which results in resulting in lower accuracy of motion intention recognition. In addition to producing sEMG signals in limb movement, muscle deformation can be produced during the muscle contraction, which is closely related to motion. Compared with sEMG, muscle deformation which is a mechanical signal has better robustness because it is not affected by skin impedance and environmental electromagnetic interference. Therefore, this project proposes a new approach to identify user’s motion intentions and achieve prosthetic control by using muscle deformation signal. .First, two new flexible sensing materials, piezoelectric electret and extensible nano-gold film, are used to collect the pressure signals and arm circumference information on the skin surface. Second, the deformation information is decoded by pattern recognition technology to identify the motion intention of the amputees. Third, Considering the limited residual muscle of amputees, the motion recognition strategy based on the fusion of muscle deformation information and sEMG signal will be developed to further enhance the accuracy and reliability of motion intention recognition. Fourth, a verification platform is developed to realize the application of muscle deformation signal in multifunctional prosthetic control. .Through this project, it is promising to develop a new way to detect human motion information and a new method of motion intention recognition, which is of great importance to improve the performance of prosthetic control.

目前的商业化假肢手主要通过表面肌电信号识别运动意图。在实际应用中，体表汗液、电极位移、电磁干扰等因素易导致表面肌电信号质量下降，造成运动意图识别准确率降低。肢体运动时，除了肌电信号，肌肉的收缩形变信息也与运动紧密相关。与肌电信号相比，肌肉形变属于机械信号，不受皮肤表面阻抗和环境电磁干扰的影响，鲁棒性较高。因此，本项目提出利用肌肉形变信号实现运动意图识别，进而应用于假肢控制。拟基于压电驻极体和可拉伸纳米金膜两种新型柔性传感材料，分别采集上肢运动时皮肤表面法向压力和手臂臂围变化信号；利用模式识别技术解码肌肉形变信息，准确获取使用者的运动意图；针对截肢者残留肌肉不足的情况，探索肌肉形变与肌电信息融合的识别策略，进一步提高运动意图识别的准确性和稳定性；开发验证平台，实现肌肉形变信号在多功能假肢手控制中的应用。本项目有望开发人体运动信息检测和运动意图识别的新方法，对提高假肢手操控性能具有重要意义。

本项目针对当前使用肌电信号进行假肢控制容易受到周围环境电磁干扰、皮肤阻抗变化、电极移动的影响导致假肢控制稳定性不足，本研究提出使用肌肉形变信号进行运动意图识别和假肢控制，针对“肌肉形变与运动意图之间的映射关系”和“基于肌肉形变信号与肌电信号的多源信息融合策略”两个关键科学问题，本研究将开展基于柔性传感的肌肉形变信号检测、肌肉形变信号分析与肌肉形变信号、肌肉形变信号与肌电信号融合的运动意图识别、多功能假肢控制验证与性能评估等四方面的研究内容，力图突破运动信息柔性传感检测、运动意图准确识别、多源运动信息融合控制等三项关键技术，实现肌肉形变信号的实时精确检测和多自由度手腕部动作准确识别，最终达到多功能假肢手的实用性和可操控性能的提高。.项目取得的研究成果包括：1） 研究高性能的传感方法和信号处理技术，实现肌肉形变信息的精确获取；2）设计先进的模式识别算法，实现截肢者运动意图的准确预测；3） 开发稳定的假肢控制方法，实现多功能假肢手3自由度的灵活抓握动作。.通过本项目研究，我们采集了多名正常受试者和前臂截肢者的动作肌肉形变信号，以及同步采集的肌电和肌肉形变多模信号。这些数据可以用来研究和开发基于人体运动多源运动信息融合的解码方法，。除了用于人体运动意图解码之外，还可以通过这些数据的分析，从电生理信号层面对正常受试者和截肢者的肌肉运动功能进行对比分析，探索截肢后的人体神经肌肉功能重塑机制。.综上所述，本项目具有较好的应用创新和实用价值，通过本项目的研究，有望探索柔性传感的新应用，开发新的运动意图识别方法和策略，提高假肢控制性能，促进多功能假肢应用。