9.1 Biosignal Processing Methods & Techniques
A Neural Network:
Recently, the field of neural networks has attracted attention as a possible method of analysis and classification of biomedical signals due to its good learning ability, adaptability and non-linear reparability networks.Hiraiwa et al. proposed a classification system of EMG signals using neural networks for EMG controlled prosthetic members. Figure Shows the schematic diagram of the EMG Classification
System, which works as follows:
The Subject makes a particular fingers movement.
Surface Electrodes detect the EMG signals when the finger movements halted then signals are averaged and analyzed by FFT.
The FFT analyzed signals are input to the network having input layer of 10 processing elements (10-PEs),hidden layer-7-PEs and output layer-5-PEs
The network matches the analyzed signals with some desired outputs (5 categories of finger movements).
The research group concluded from experiments that the ability of neural networks to recognize EMG patterns is superior to conventional methods using linear separable functions.
B.Short Time Fourier Transform
Fourier Transform(FT) is one of the most widely used methods of signal processing.FT method is important to Signal Analysis, Since they provide a tool to relate time with frequency contents of a given signal. The FT method has a number of disadvantages when it comes to analyze biomedical signals. A major disadvantage is that FT methods are not appropriate for non-stationary signals or signals with short live components such as biomedical signals, therefore short time Fourier transform has been introduced. In STFT,a window function with fixed width is chosen and then this window is slid throughout the whole signal. In this case, the signal inside the window is stationary. Then the inner product of the signal x(t),inside the window with a discrete family of integrable functions hm,n,(t) is computed using the following equation
However, the use of a fixed length window means that the resolution of the spectrogram, in both time and frequency is fixed. Because of that Wavelets Transform (WT) is introduced.
C. Wavelets Transform
The wavelets Transform (WT) is signal decomposition on a set of basis functions, obtained by dilations, contractions and shifts of a unique function, the wavelet prototype. The Wavelets Transform has many similarities with the STFT but is fundamentally different in that its wavelet prototypes are not of fixed length. In other words, it is narrow at high frequencies and broad at low frequencies. This enables WT to “zoom-in” on the short-lived components of the biomedical signal to look at finer details.Sparto et al used WT and STFT methods for the measurement of the frequency content of trunk muscle electromyography (EMG) to quantify the amount of fatigue endured by workers during industrial tasks, as well as a tool that may guide the training and rehabilitation of healthy and injured workers.
9.2 LABVIEW SIGNAL ACQUISITION:
Basic Block Diagram:
Band Stop Filter:
Band Pass Filter:
EMG Waveforms Recorded: