| Humans have a massive activity system composed muscles, joints, bones and nerves. This system helps humans to perform numerous complex tasks. Muscles are the primary component in constructing a motion of the human body. Muscles, as any other parts of the human body, receive its commands from the brain as electrical signal that flux through nerves. This electrical signal commands muscles to contracts in a harmonic way. Any shortcoming in the muscle system affects the abilities of the human body’s motion and tasks. The primary role that muscles play in humans motions turned researches attentions to study muscles and their functionalities. Muscle force and fatigue are two of the important muscle functionalities. These metrics gain insights on humans’abilities and weakness. However, how these metrics can be measured or estimated.The emerged of electromyography (EMG) signal allowed researches to record the elec-trical signals that flux in muscles. This concept became the base stone in measuring muscle activities. Nevertheless, processing the recoded EMG signal to extract muscle force and fatigue revealed and encountered many issues and problems. This work attempts to solve three of these problems.First, the process of extracting muscle force is based on the two converting steps. The first step extracts muscle activation from EMG signal. In the second step, Hill-type model is applied to calculate the muscle force. This process is complex that requires the optimization of many coefficients and values. In addition, many equations need to be solved. We pro-posed two new muscle force extraction models that simplify the extraction method. The first model utilizes the neural network (ANN) theory. Fuzzy logic is the base of the second one. We have conducted a measurement experiments on six human arms’muscles. To facilitate the implementation of our models, MATLAB was used. We compared our two new models according to their calculation speed and stability. Our investigation shows that fuzzy model is faster and more stable than neural network model. However, it was more complex than the ANN model.Second, EMG recording and processing is the fundamental step beneath all muscle activities studies. Measuring EMG signal is liable for a large amount of noise. To extract a clean and clear signal, signal processing techniques are used. The signal processing of EMG signal consists of three steps. Nevertheless, they are easy to be deployed. However, the output of this process may reduce the accuracy of the recorded data. To neglect the signal processing techniques, we have proposed a genetic algorithm (GA) technique. This technique will be fed with raw EMG signal. The output of this model is a signal that can be utilized in muscle force and fatigue extraction. To demonstrate the performance of this model, first, we have implemented a new ANN muscle force extractor with GA beneath it. Second, we have compared the output extracted force with the extracted force of our original ANN muscle force estimator. Our results show the similarity and the convergence between the two models. Finally, muscle fatigue which is defined as the muscle power descending factor is one of the important risks that influence muscles. To estimate muscle fatigue, many methods and techniques have been proposed. However, as in muscle force extraction techniques, muscle fatigue techniques are complex. We proposed a new method to estimate muscle fatigue based on fuzzy logic theory. The new model facilitates muscle fatigue estimation in one hand and can be generally utilized for any arm’s muscle on the other hand. This model uses the raw EMG signal without any signal processing steps as in our previous model. However, it utilizes fuzzy logic rather than GA. To demonstrate its capability, we conducted a measurement experiment on fifty volunteers. We measured the EMG signal that commands two muscles, namely, biceps and triceps brachii. In addition, we have investigated the relation between muscle fatigue and the human age. |
PDF Full Text Request