基于脑血氧信息的运动状态识别及自适应助行轨迹规划方法研究

Faced the requirements of patients with motor dysfunction for realizing independent walking, intelligent driving of walking-assistive devices such as exoskeletons has attracted much attention. And the classification of motion state is one of the key problems to be solved. As for the patients who lost important muscle groups or have weakened muscle functions, the methods of identifying motion state based on EMG signals or biomechanics information have great limitations. In order to increase the scope of the patients, the project proposed a method of classifying motion state based on hemoglobin information in cerebral cortex, and a method of planning adaptive motion trajectory based on the desired motion state (step length and walking velocity). Characteristics of brain functional network during motor imagery are analyzed and a novel method for feature extraction is proposed by using the original neural information. The influence of walking velocity and step length on the characteristics of blood oxygen is studied, and a coupling model of hemoglobin characteristics is constructed based on two-dimensional variables of walking velocity and step length. The aim is to identify the desired state of walking velocity and step length of the subject simultaneously according to typical feature vectors in two-dimensional space. At the same time, the characteristics of gait parameters in the different motion state are studied, and mathematical models to describe joint movement are established. The aim is to present an adaptive motion trajectory for knee with the desired step length and walking velocity, and then to give the controller of walking-assistive devices as a reference movement. The project seeks breakthroughs of key problems of science and technology in the research on recognition of motion state and planning of movement trajectory. It lay the foundation for realizing intelligent control of a walking-assistive device, thus, it has important academic significance and social significance.

针对有运动功能障碍的患者实现独立行走的迫切需求，外骨骼等助行设备的智能驱动备受关注，其中运动状态的判别是要解决的关键问题之一。基于肢体生肌电信息或运动信息识别运动状态的方法，对于关键肌肉群缺失或肌肉功能减弱的患者存在很大的局限性。为增加患者适用范围，本项目提出基于脑血氧信息的运动状态识别及自适应运动轨迹规划的方法研究。应用最原始的神经元信息，分析运动想象期间的大脑功能网络特性并提出特征提取新方法，研究步速和步长对血氧特征的影响规律，建立基于步速和步长二维变量的血氧特征耦合模型，并基于典型特征向量判别期望实现的步速和步长状态。同时，研究不同运动状态下的步态参数特征并建立关节运动数学模型，基于步速和步长状态同步描述关节运动轨迹，为助行设备的控制提供自适应参考运动。本项目寻求运动状态识别与运动轨迹规划研究中关键科学问题与技术问题的突破，为助行设备的智能驱动奠定基础，具有重要的学术意义和社会意义。

针对有运动功能障碍的患者实现独立行走的迫切需求，外骨骼等助行设备的智能驱动备受关注，其中运动状态的判别是要解决的关键问题之一。项目提出基于脑血氧信息的连续运动状态识别方法研究。项目研究不同步速和不同步长状态下对应脑血氧信息的特征规律，基于典型血氧特征向量判别期望实现的步速和步长状态，三种步长识别率76.67%（步速不同） ，三种步速识别率75%（步长不同）。同时，通过深度挖掘脑血氧时频特征和脑功能网络动态特征等，分别建立行走启停意图、步态调整意图和步态调整状态的动态识别模型，行走意图的识别率达到98.75%，平均延时0.57秒；步态调整意图的识别率达到91.25%，平均超前判别0.12秒；步态调整状态的判别平均识别率达到83.75%。项目研究提出的解码方法可为助行器械实时提供控制指令，为实现基于脑机接口技术的助行器械智能驱动奠定理论基础与技术基础，具有重要的学术意义和社会意义。