面向医疗康复应用的穿戴式步态分析技术研究

Human gait contains complex kinematics and dynamics information as well as abundant physiological, psychological and pathological information. Under the construction of healthy China, gait analysis is attracting increasing research interests in the fields of medicine and engineering. However, gait analysis in clinical rehabilitation practice still mainly relies on the doctors' clinical observation, and the industry-university-research transformation has not been effectively established yet. Aiming at the feature extraction and analysis of human gait data, this project will focus on the low cost, personalized, wearable gait analysis technology for medical rehabilitation applications, by combining the gait analysis technology with wearable sensing technology, data fusion technology, pattern recognition technology, error correction technology, and nonlinear analysis techniques. An adaptive gait recognition algorithm for pathological gait will be proposed based on multimodal data fusion and lower limb kinematics model, so as to perform a comprehensive, objective, accurate, consistent, and quantitative evaluation of gait patterns, and to reveal the key points and deep factors in the process of gait abnormalization. Furthermore, to aid in clinical decision-making, an intelligent diagnostic algorithm for pathological gait will be developed based on significant feature selection and nonlinear analysis technology, so as to provide the scientific basis and technical support for fall risk management, pathogenesis research, disease prediction, disease diagnosis, treatment planning and rehabilitation evaluation, which has significant social and economic significance.

人体步态蕴含复杂的运动学和动力学信息以及丰富的生理、心理和病理信息，在健康中国建设的时代背景下，步态分析成为了医工结合的研究热点。然而，目前医疗康复中的步态分析，依然主要依靠医师的临床观察，产学研转化链尚未有效形成。本项目旨在研究面向医疗康复应用的低成本、个人化、穿戴式步态分析技术，以人体步态数据的特征提取及分析为核心，通过将步态分析技术与可穿戴传感技术、数据融合技术、模式识别技术、误差校正技术、非线性分析技术等有机结合，研究融合多模态步态信息与下肢运动学模型的适用于病理步态的特征识别算法，对病理步态进行全面、客观、精确、一致的量化评估，以揭示步态异常的关键环节和深层因素，研究集成显著性特征选取与非线性分析技术的病理步态智能诊断算法以辅助临床决策，为跌倒风险管理、发病机理研究、疾病早期预测、特定疾病诊断、治疗方案制定、康复效果评价等提供科学依据和技术支持，具有显著的社会意义和经济意义。

人民健康水平的现状和人口老龄化的程度，为健康中国建设带来严峻挑战的同时，也为医工结合研究创造了前所未有的机遇。智慧城市、智慧医疗、智慧养老等理念的实践，均离不开人体运动信息的精准监测。本项目借助于体感网和穿戴式监测技术，以人体最基本的行走运动为抓手，通过医工结合为人体健康的定量监测和精准干预提供科学依据和技术支持。经过三年的潜心研究，项目组按照研究计划完成了研究目标，开展了三方面的研究工作：人体运动数据分割与步态识别、步态参数解算与参数误差校正、步态特征评定与临床实验验证。提出了一种融合多模态步态信息与下肢运动学模型的适用于病理步态的特征识别算法，提出了一种集成显著性特征选取与非线性分析技术的典型病理性步态的智能诊断算法，研制了一套能够客观量化用户下肢步态信息的穿戴式步态分析系统。本项目研究成果除了应用于医疗康复外，也可应用于动作识别、身份识别、义肢开发、行人定位与导航、康复机器人、外骨骼机器人、足式机器人甚至陆地动物的步态分析等诸多领域，具有重要的社会意义和经济意义。

内容获取失败，请点击重试