SEMG-based shoulder-elbow composite motion pattern recognition and control methods for upper limb rehabilitation robot. Issue 3 (16th October 2018)

• Record Type: Journal Article
• Authors: Zhang, Xiufeng; Dai, Jitao; Li, Xia; Li, Huizi; Fu, Huiqun; Pan, Guoxin; Zhang, Ning; Yang, Rong; Xu, Jianguang
• Is Part Of: Assembly automation. Volume 39:Issue 3(2019)
• Page Start: 394
• Page End: 400
• Publication Date: 2018-10-16
• Subjects: BP network -- Support vector machine (SVM) -- Composite motion pattern recognition of shoulder-elbow -- Upper limb rehabilitation training system; Automation -- Periodicals; Automatic machinery -- Periodicals; Assembly-line methods -- Periodicals; Industrial engineering -- Periodicals; 670.42705
• DOI: 10.1108/AA-11-2017-148
• Languages: English
• ISSNs: 0144-5154

Abstract

Abstract : Purpose: This paper aims to develop a signal acquisition system of surface electromyography (sEMG) and use the characteristics of (sEMG) signal to interference action pattern. Design/methodology/approach: This paper proposes a fusion method based on combining the coefficient of AR model and wavelet coefficient. It improves the recognition rate of the target action. To overcome the slow convergence speed and local optimum in standard BP network, the study presents a BP algorithm which combine with LM algorithm and PSO algorithm, and it improves the convergence speed and the recognition rate of the target action. Findings: Experiments verify the effectiveness of the system from two aspects the target motion recognition rate and the corresponding reaction speed of the robotic system. Originality/value: The study developed a signal acquisition system of sEMG and used the characteristics of (sEMG) signal to interference action pattern. The myoelectricity integral values are presented to determine the starting point and end point of target movement, which is more effective than using single sample point amplitude method.