EMG BASED CONTROL OF WRIST EXOSKELETON

Abstract

The significance of human motion intentions in a designed exoskeleton wrist control hand is essential for stroke survivors, thus making EMG signals an integral part of the overall system is critically important. However, EMG is a nonlinear signal that is easily influenced by several errors from its surroundings and certain of its applications require close monitoring to provide decent outcomes. Hence, this paper proposes to establish the relationship between EMG signals and wrist joint angle to estimate the desired wrist velocity. Fuzzy logic has been selected to form a dynamic modelling of wrist movement for a single muscle at different MVC levels and double muscles at a similar MVC level. The physical model of the exoskeleton hand using Simmechanics Matlab software has been developed to validate the performance of the fuzzy logic output result from both dynamic modelling approaches. A PID controller has been developed to smooth the exoskeleton hand movement fluctuations caused by the fuzzy logic decision-making process. As a conclusion, results showed a strong relationship between EMG signals and wrist joint angle improved the estimation results of desired wrist velocity for both dynamic modelling approaches hence strengthened the prediction process by providing a myoelectronic control device for the exoskeleton hand. ABSTRAK: Kepentingan dalam mengetahui kehendak gerakan pergelangan tangan manusia adalah penting untuk pesakit strok yang terselamat, justeru menjadikan isyarat EMG amat penting pada keseluruhan sistem. Walau bagaimanapun, EMG adalah isyarat tidak linear yang mudah dipengaruhi ralat sekitaran dan memerlukan pemantauan rapi bagi hasil yang baik. Oleh itu, kajian ini mencadangkan kewujudan hubungan antara isyarat EMG dan sudut sendi pergelangan tangan bagi menganggarkan halaju pergelangan tangan yang dikehendaki. Logik kabur (fuzzy logic) telah dipilih bagi membentuk model dinamik pergerakan pergelangan tangan pada otot tunggal di tahap MVC yang berbeza dan otot berganda pada tahap MVC yang serupa. Model fizikal rangka luar tangan menggunakan perisian Matlab Simmekanik telah dibangunkan bagi mengesahkan prestasi Logik Kabur daripada kedua-dua pendekatan model dinamik. Pengawal PID telah dibangunkan bagi melicinkan gerakan turun naik tangan yang disebabkan proses membuat keputusan oleh Logik Kabur. Sebagai kesimpulan, dapatan kajian menunjukkan hubungan yang kukuh antara isyarat EMG dan sudut sendi pergelangan tangan. Ini meningkatkan anggaran dapatan halaju pergelangan tangan yang dikehendaki bagi kedua-dua pendekatan model dinamik seterusnya mengukuhkan proses ramalan melalui peranti kawalan mioelektronik rangka tangan.

Keywords

• Myoelectronic
• Control devices
• Fuzzy Logic
• PID controller
• Exoskeleton wrist design