Signal processing Application

This was research based experiment wherein we studied an application of dsp processor.We formed a group of 5 students. The members were Nikita Pagar, Golappagouda Patil, Anushree Mhatre, Apoorva Raut and Sahil Rai. The topic we selected was “Detection and Processing of Electromyography signals”.

the paper name   

Electromyogram Amplitude Estimation with  Adaptive Smoothing Window Length

Abstract—Typical electromyogram (EMG) amplitude estimators use a fixed window length for smoothing the amplitude estimate. When the EMG amplitude is dynamic, previous research suggests that varying the smoothing length as a function of time may improve amplitude estimation. This paper develops optimal time-varying selection of the smoothing window length using a stochastic model of the EMG signal. Optimal selection is a function of the EMG amplitude and its derivatives. Simulation studies, in which EMG amplitude was changed randomly, found that the “best” adaptive filter performed as well as the “best” fixed-length filter. Experimental studies found the advantages of the adaptive processor to be situation dependent. Subjects used real-time EMG amplitude estimates to track a randomly-moving target. Perhaps due to task difficulty, no differences in adaptive versus fixed-length processors were observed when the target speed was fast. When the target speed was slow, the experimental results were consistent with the simulation predictions. When the target moved between two constant levels, the adaptive processor responded rapidly to the target level transitions and had low variance while the target dwelled on a level.

the patent   

EMG control of prosthesis

Inventors: RISO, Ronald R. (US/UK): Herulf Trolles Gade 28, 2.th., DK-9000, Aalborg (UK).

Application Number: PLT/DK00/00464

Filling Date: 21 August 2000

International Patent Classification: A61B

https://drive.google.com/drive/folders/0BxILgHTXcwaNYzBfNmlpS2VuWUE

OSM/OAM

Overlap Add Method (OAM) and Overlap Save Method (OSM) are efficient methods to calculate the convolution between long length input signal and finite impulse signal. 

we took a 15pt long sequence as input and processed it. We used FFT to implement Overlap add method. The long sequence was converted into smaller finite lenght sequences and then processed.

We learned that FFT is used due to its speed.

Also OAM is a practical method to process real-time infinite lenght sequences.

https://drive.google.com/drive/folders/0BxILgHTXcwaNYzBfNmlpS2VuWUE

Basic operations using DSP processor

The demonstration of DSP processor TMS320F28375 kit was given by JaiGanesh (EXTC). He demonstrated the basic operations such as, addition, subtraction, multiplication , division, shift and rotate operations etc.. using DSP processor. We verified the contents of register before and after execution.

Digital FIR filter design using frequency sampling method

The frequency sampling method allows us to design recursive and nonrecursive FIR filters for both standard frequency selective and filters with arbitrary frequency response. The frequency response of LPF and HPF are plotted using frequency sampling method. Ripples in the stopband are obtained by decreasing amplitude. The phase plot is linear and similar for both LPF and HPF if order of the two filters is same. Since phase is linear, output is not distorted.

In this the desired frequency response Hd(w) is sampled at w=(2*pi*k)/N and the frequency samples thus obtained are taken as DFT coefficients. FIR filter with impulse response is then calculated by IDFT.

Thus, for the same values of attenuation in stop band and pass band, pass band and stop band frequencies and sampling frequencies we observe that the order of the FIR filter is much higher compared to that of the IIR filter.

https://drive.google.com/drive/folders/0BxILgHTXcwaNYzBfNmlpS2VuWUE

Digital FIR filter designing using windowing method

Different types of windows are used to design FIR filter. Here, we have used hanning window to design FIR filter. FIR filter has linear phase and also the observed values of As and Ap from magnitude spectrum are closed to the values entered. 

In this experiment we had to find the output of a Linear Phase FIR filter using Window Fucntion. A window function is required to truncate infinite samples of hd(n). The window function was decided on the basis of the values of attenuation in stop band. Thus, we performed the experiment for window Bartlett function.

As=21 for Rectangular

25 for Bartlett

44 for Hanning

53 for Hamming

74 for Blackmann

We learnt that Hanning Window function gives more attenuation in stop band than Bartlett window, hence it is a better window function.

https://drive.google.com/drive/folders/0BxILgHTXcwaNYzBfNmlpS2VuWUE

Design of Chebyshev Filter

Chebyshev I filter has ripples in the passband and no ripple in the stop band. Cheebyshev filter II has ripples in stopband and no ripple in passband. 

We designed chebyshev filter using scilab. Plotted the magnitude spectrum and verified the values of As and Ap in stopband and passband. Those values were approximately same. All poles were located inside the unit circle.

The order of the filter was calculated using the given values of attenuation in pass band and stop band, pass band and stop band frequencies and sampling frequency. The normalised and denormalised H(s) was calculated, from which the Transfer Function H(z) was calculated.

We observed and plotted the graphs and learnt that:

1. In the magnitude response of the Low Pass Filter, there is ripple in the pass band whereas the stop band is monotonic and there is no ripple.

2. In the magnitude response of the High Pass, there is ripple in the stop band whereas the response is monotonic in the pass band.

https://drive.google.com/drive/folders/0BxILgHTXcwaNYzBfNmlpS2VuWUE

Convolution and Correlation Algorithm

In the circular convolution aliasing takes place where the last few values get aliased with ones in the start. In circular convolution the length of the output convoluted signal is , N= max(L,M)

where, L= length of first input signal

M= Length of second input signal.

for linear convolution, N=L+M-1

https://drive.google.com/drive/folders/0BxILgHTXcwaNYzBfNmlpS2VuWUE