Exoskeleton nursing robot is a typical human-machine co-drive system. To full play the subjective control and action orientation of human, it is necessary to comprehensively analyze exoskeleton wearer's surface electromyography (EMG) in the process of moving patients, especially identifying the spatial distribution and internal relationship of the EMG information. Aiming at the location of electrodes and internal relation between EMG channels, the complex muscle system at the upper limb was abstracted as a muscle functional network. Firstly, the correlation characteristics were analyzed among EMG channels of the upper limb using the mutual information method, so that the muscle function network was established. Secondly, by calculating the characteristic index of network node, the features of muscle function network were analyzed for different movements. Finally, the node contraction method was applied to determine the key muscle group that reflected the intention of wearer's movement, and the characteristics of muscle function network were analyzed in each stage of moving patients. Experimental results showed that the location of the myoelectric collection could be determined quickly and efficiently, and also various stages of the moving process could effectively be distinguished using the muscle functional network with the key muscle groups. This study provides new ideas and methods to decode the relationship between neural controls of upper limb and physical motion.
外骨骼护理机器人是典型的人机共驱动系统，为了充分发挥人的主观控制能力与动作导向能力，需要全面分析外骨骼穿戴者在搬运患者过程中表面肌电信息的空间分布情况和内在联系。针对肌电电极位置选择以及各通道间的内在关系等问题，将上肢肌群复杂系统抽象为肌肉功能网络。首先，利用互信息分析上肢各肌电通道间的关联特性，构建肌肉功能网络；其次，通过计算网络的节点特性指标，分析不同运动过程中肌肉功能网络特点；最后，利用节点收缩法确定反映用户运动意图的关键肌群，分析搬运过程各阶段的肌肉功能网络特性。实验证明，该方法可以快速有效地确定肌电采集位置，关键肌群的肌肉功能网络特性可以有效区分搬运过程的各阶段，为上肢神经控制与身体运动的关系解码提供了新思路和新方法。.
Keywords: exoskeleton nursing robot; muscle functional network; node contraction method; surface electromyography; upper extremity muscles.