مسار بولونيا
EN

محمد زكي حسين جاسم علي الفائز

السنة عنوان البحث نشر البحث
2018 Humanoid Robotic Hand (HRH) Based on EMG signal for Amputees Persons International Journal of Emerging Trends in Engineering Research
his paper proposes a system that designed and implemente o be used for persons who lost their limbs because o ccident, wars or any other diseases. The idea of this paper i o design and implement Humanoid Robotic Hand (HRH) he HRH was built by using 3D printer technology of har olylactide (PLA) filament. The different components of th ands were separately built and then assembled, which give asy manufacturing at the same time great latitude to choos materials, and also built by six servo motors while five serv motors used to move the fingers and the sixth servo moto sed to rotate the wrist. The designed HRH system wa ontrolled by electromyography (EMG) signals were utilize o classify seven classes of movements in offline mode an ive movements in real-time. The EMG signals wer measured by using three surface electromyography (sEMG MyoWare muscle sensors, which were located on the forear n three muscles (Extensor Carpi Ulna, Extensor Carp adius and Extensor Carpi Digitorum) and use Arduin Mega microcontroller as an analogue to digital converter t ake the signal from the sensor and also use data collector t ontrol the humanoid robotic hand. The proposed patter ecognition system was investigated in an offline mode t nhance it and to develop the classification accuracy of th ystem by using the (Integral Absolute Value (IAV), Mea bsolute Value (MAV) , Root Mean Square (RMS) Waveform length (WL), Zero Crossing (ZC), Slope Sig hange (SSC) and Autoregressive (AR) as feature extractio Principal Component Analysis (PCA) as feature reduction -Nearest Neighbor (k-NN) and Linear Discriminan nalysis (LDA) algorithms as classifiers. Furthermore, th ffects of electrodes’ position on the forearm and the numbe f channels on the efficiency of the pattern recognition syste ere investigated too. The results showed that th erformance of the LDA is better than the k-NN because th ccuracy of LDA is 91.1056% and the accuracy of k-NN i 7.5849% these percentages are in the offline mode and i eal time mode the accuracy is 84% when using LD