In the relaxed state of muscle, a more or less noise-free EMG baseline is observed in EMG plot. The raw EMG baseline noise depends on many factors, especially the quality of the EMG amplifier, the environmental noise and the condition of subject under investigation. In case a state of the art amplifier is available and skin is properly prepared, the averaged baseline noise should not be higher than 3 to 5 microvolts, 1 to 2 achievable. The investigation of the EMG baseline quality is a very important checkpoint of every EMG measurement. Interfering noise or problems within the detection apparatus may results in increased base activity or muscle.
The healthy relaxed muscle shows no significant EMG activity due to lack of depolarization and action potentials. The raw EMG spikes are usually random in shape, which means one raw recording burst cannot be precisely reproduced in exact shape. This is due to the fact that the actual set of recruiting motor units constantly changes within the matrix/diameter of available motor units. If suddenly two or more motor units are fired at the same time and they are located near the electrodes, a strong superposition spike is produced. The application of a smoothing algorithm (e.g. moving average filter) or selecting a proper amplitude parameter (e.g. area under the rectified curve), eliminates non- reproducible contents of the signal.
A raw SEMG signal ranges between ±5000 μV and typically frequency contents 6 – 500 Hz and the most of frequency power is between 20 and 150 Hz.
Besides basic physiological and biomechanical studies, SEMG is established as an evaluation tool for applied research, physiotherapy/rehabilitation, sports training and interactions of the human body to industrial products and work conditions is as shown in Fig. 1.1.
Fig. 1.2 Block diagram of SEMG research fields
In 1940‟s an application of EMG signal was deployed for prosthesis control. The prosthesis is of many types but EMG signal is used as input for the control of power prosthesis. The signal is used to select and modulate a function of a multifunction prosthesis [1]. ELECTROMYOGRAPHY Medical Research Orthopedic Surgery functional Neurology Ergonomics Analysis of demand Risk Prevention Ergonomics Design Rehabilitation Post surgery/accident physical Therapy Prosthesis Sports Science Biomechanics Movement Analysis Athletes Strength Training
1.1 Prosthesis
Prosthesis is artificial substitute for a missing part of the body. The artificial parts that are most commonly thought of as prostheses are those that replace lost arms and legs, but bone, artery, heart valve replacements are common and artificial eyes and teeth are correctly termed prostheses.
The medical specialty that deals with prostheses is called prosthetics. The origin of prosthetics as a science is attributed to the 16th-century French surgeon Ambroise Pare [2]. Later workers developed upper-extremity replacements, including metal hands made either in one piece or with movable parts. The solid metal hand of the 16th and 17th centuries later gave way in great measure to a single hook or a leather-covered, nonfunctioning hand attached to the forearm by a leather or wooden shell .Improvement in the design of prostheses and increased acceptance of their use have accompanied major wars. New lightweight materials and better mechanical joints were introduced after World Wars I and II.
The different prostheses developed by the main prosthetic societies: UTAH, OTTA BOCK and PROTEOR concentrate 90% of the market. These are classified in three categories namely; aesthetic prostheses, body-powered prostheses and myoelectric prostheses.
1.1.1 Aesthetics Prosthesis
This type of prosthesis is generally used by patients and their aim is only aesthetics. The prosthetic part is created from a standard mold and resemblance to the healthy member. This kind of prosthesis does not carry out any movement. It only serves to restore the patient’s body appearance. This kind of prosthesis is far instance manufactured by the OTTO BOCK society.
1.1.2 Body-powered Prostheses
A body-powered prosthesis, also known as conventional or cable-driven prosthesis, is powered and controlled by gross body movements. This movement usually of the shoulder, upper arm, or chest is captured by a harness system and used to pull a cable that is connected to a TD (hook or hand). For some levels of amputation or deficiency, an elbow system can be added to provide additional motion. For a patient to control a body-powered prosthesis, the individual must be capable of producing at least one or more of the following gross body movements:
1.1.3 Myoelectric Prostheses
Myoelectric signals (Electromyogram or EMG) are electrical signals that are registered for muscular activities. A large number of applications are possible with these signals. The functional motor activities can be measured by placing the surface electrodes directly on the skin. The EMG signals are complex with noise and they are easily influenced by many factors. The EMG signal requires several treatments before it can be interpreted and used. UTAH society was the first to propose the EMG technology to control the prosthesis. The OTTO BOCK society also proposes prosthesis of hand coupled with a myoelectric elbow. Unfortunately, the whole system proposed by this society is too expensive for the patient. The hand is a tree legs grip with an aesthetic glove [3].
1.2 Algorithms for EMG Classification
The control of assistive devices and exoskeletons using EMG signals has been the focus of many researchers. As EMG signals are complex in nature, EMG classification for motion detection is a challenging task. The various approaches used to efficiently classify the EMG signals are summarized as follows (1) Neural network (2) Fuzzy logic (3) Hybrid fuzzy-neural approaches and (4) Particle swarm optimization-SVM based.
In presented work fuzzy logic approach are used for classification of EMG signals. There has been many works on fuzzy approaches to EMG classification. Fuzzy logic has the ability to deal with imprecise, uncertain and imperfect information. The strength of fuzzy logic lies in the fact that it is based on the reasoning inspired by human decision-making. This fuzzy logic is used to handle the vagueness intrinsic to many problems by representing them mathematically. Fuzzy logic systems are advantageous in biomedical signal processing and classification. Biomedical signals are not always strictly repeatable, and may sometimes even be contradictory. One of the most useful properties of fuzzy logic systems is that contradiction in the data can be tolerated. Furthermore using trainable fuzzy systems, it is possible to discover patterns in data, which are not easily detected by other methods, as, can also be done with neural network. Other neural network approaches are discussed in literature work.
In the presented work the objective of EMG signal classification is achieved in a step wise approach. The signals from subjects are acquired using single channel Biokit system. The acquired signals are processed using LabVIEW for feature extraction. The signal processing performed involves signal amplification, signal filtration and sampling using analog to digital converter.
To classify processed EMG signal, a fuzzy logic based classifier is developed by LabVIEW 2012 (evaluation version). The performance of developed classifier is evaluated using available test data acquired from subjects in lab. The test data is obtained from two points (1) flexor carpum ulnaris (below elbow) for gripping and (2) biceps brachhi (between elbow and shoulder) for lifting movements of different weights.
The presented work organized in 6 chapters. First chapter is an introduction of EMG signals, EMG signals for prosthesis and algorithms of classification in the field of EMG signal classification. Chapter 2 consists of literature review, to study the EMG signal analysis and classification. It presents some time domain and frequency domain feature extraction techniques of EMG signal. It provides guideline to design the fuzzy classifier for this application. Chapter 3 summarized the material and methods used for acquisition of signal and classification. Chapter 4 consists of the solution methodology, Setup of signal acquisition, feature extraction and design classifier for EMG signal. In chapter 5 results are presented and designed classifier is tested with 180 EMG signals so that the success rate of the classifier can calculate. Chapter 6 concludes the presented work and future scope is summarized.