EMG motion pattern classification through design and optimization of Neural Network

Design and Optimization of Levenberg-Marquardt based Neural Network Classifier for EMG Signals to Identify Hand Motions

“…The EMG signals are obtained for different kinds of hand motions, which are further denoised and processed to extract the features. Extracted time and time-frequency based feature sets are used to train the neural network [14]. The block diagram of ANN based EMG signal classification is shown in Fig.1 [15]- [18], are extracted from the EMG signals and used as inputs to the neural network.…”

“…The training and performance testing of ANN are done by applying both Levenberg-Marquardt (trainlm) and scale conjugate gradient (trainscg) algorithms. During each time of the training period, both algorithms function to adjust the weights and biases of the network in such a way as to minimize the MSE and hence increase the rate of network performance [14].…”

Predicting EMG Based Elbow Joint Torque Model Using Multiple Input ANN Neurons for Arm Rehabilitation

“…However there is constraint in determining the number of neurons. If the numbers of hidden neurons is too large, the network requires more memory and the network become more complicated while if the number of hidden neurons is too small, the network would face difficulty to adjust the weigh properly and could cause over fitting which is problem where the network cannot be generalized with slightly different inputs [10]. The input signal is propagated forward through network layer using back propagation algorithm.…”

“…It also has output data which is torque of correspondent arm motion. The training process was iteratively adjusted to minimize the error and increased the rate of network performance [10]. MATLAB software is used to construct the BPNN network.…”

LLE-ELM Recognition Method for the Joint Range of Shoulder and Elbow Motion

“…Some international rating scales evaluate disability degree for upper limb according to the finished state of some movements, such as activity of daily living scale (ADL) [2], Fugl-meyer motor assessment (FMA) [3] and etc. However, most researchers used standard supervised machine learning algorithms and focused on estimating the forearm, wrist and hand [4][5][6][7]. These approaches did not consider the shoulder and elbow joint angles that are not enough to cover rehabilitation training and assessment of the upper limb, and lack of relatively satisfied recognition accuracy and speed in sEMG signals.…”