Identification of EMG signals using discriminant analysis and SVM classifier Ahmet Alkan ⇑ , Mücahid Günay Kahramanmaras Sutcu Imam University, Department of Electrical & Electronics Engineering, Kahramanmaras, Turkey article info Keywords: EMG Discriminant analysis Cross validation SVM abstract The electromyography (EMG) signal is a bioelectrical signal variation, generated in muscles during volun- tary or involuntary muscle activities. The muscle activities such as contraction or relaxation are always controlled by the nervous system. The EMG signal is a complicated biomedical signal due to anatomi- cal/physiological properties of the muscles and its noisy environment. In this paper, a classification tech- nique is proposed to classify signals required for a prosperous arm prosthesis control by using surface EMG signals. This work uses recorded EMG signals generated by biceps and triceps muscles for four dif- ferent movements. Each signal has one single pattern and it is essential to separate and classify these pat- terns properly. Discriminant analysis and support vector machine (SVM) classifier have been used to classify four different arm movement signals. Prior to classification, proper feature vectors are derived from the signal. The feature vectors are generated by using mean absolute value (MAV). These feature vectors are provided as inputs to the identification/classification system. Discriminant analysis using five different approaches, classification accuracy rates achieved from very good (98%) to poor (96%) by using 10-fold cross validation. SVM classifier gives a very good average accuracy rate (99%) for four movements with the classification error rate 1%. Correct classification rates of the applied techniques are very high which can be used to classify EMG signals for prosperous arm prosthesis control studies. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Bioelectrical signals are usually taken to be electric currents produced by the sum of electrical potential differences across a specialized tissue or organ as a result of different electrochemical events happening in the body. EMG signal is one of the best-known bioelectrical signals which can be detected over the skin surface and are generated by the electrical activity of the muscle fibres during contraction or relaxation. Each movement of muscles corresponds to a specific pattern of activation of several muscle fibres; therefore multi-channel EMG recordings can be used to identify the movement. Due to the com- plex nature of the signal, detailed analysis and classification is of- ten difficult, especially if the EMG relates to movement (Kumar, Ma, & Burton, 2001). For this purpose, different pattern recognition schemes, consist- ing of feature extraction and classification, have been applied (Par- ker, Englehart, & Hudgins 2004). This concept has been used for the development of myoelectric prosthesis control systems obtained by classification of EMG signals (Choi & Kim, 2007; Englehart, Hud- gins, & Parker, 2001; Hu & Nenov, 2004; Kumar et al., 2001; Lucas, Gaufriau, Pascual, Doncarli, & Farina, 2008; Parker & Scott, 1986; Parker et al., 2004; Tscharner, 2000; Wojtczak, Amaral, Dias, Wol- czowski, & Kurzynski 2009. As in Lucas et al. (2008), the discrete wavelet transforms (DWT) based representation space is used for supervised classification of multi-channel surface electromyography signals with the aim of controlling myoelectric prostheses. They applied a support vector machine (SVM) approach to classify a multichannel representation space. They optimized the mother wavelet with the criterion of min- imum classification error, as estimated from the learning signal set. Then the method was applied to the classification of six hand move- ments with recording of the surface EMG from eight locations over the forearm. For all subjects using the eight channels they reported a misclassification rate as (mean ± S.D.) 4.7 ± 3.7% with the proposed approach while it was 11.1 ± 10.0% without proposed technique. They stated that DWT and SVM can be implemented with fast algo- rithms and their method is suitable for real-time applications. The success of a myoelectric control scheme depends largely on the classification accuracy. Englehart et al. (2001), proposed a no- vel approach that demonstrates greater accuracy than their previ- ous work. They used wavelet-based feature set, reduced in dimension by principal components analysis. Further, it is exposed that four channels of myoelectric data increases the classification accuracy, as compared to one or two channels. They claimed that a robust online classifier is constructed, which produces class deci- sions on a continuous stream of data. Choi and Kim (2007) investigated to design an assistive real- time system for the upper limb disabled to access a computer via 0957-4174/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2011.06.043 ⇑ Corresponding author. E-mail addresses: aalkan@ksu.edu.tr (A. Alkan), mucahidgunay@gmail.com (M. Günay). Expert Systems with Applications 39 (2012) 44–47 Contents lists available at ScienceDirect Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa