ELE8059 Digital Signal Processing Coursework Assignment 2: Inter-Peak Interval and Peak Amplitude Estimation from Noisy Signal Assignment Note As with assignment 1, the coursework is strictly individual, and students should not collaborate on the assignment. Assignment Brief In this assignment, you are set a practical task of estimating the inter-peak interval and the amplitudes of the primary and secondary peak of an ECG-like, periodic measurement signal that is heavily contaminated by noise as well as by a 50 Hz mains electricity component. The underlying, uncontaminated signal is periodic, with one period of the form displayed in Fig 1. Figure 1: One period of an example underlying signal, characterised by the inter-peak interval ( ), the primary-peak ampltude (a), and secondary-peak amplitude (b). The assignment is to: a) perform frequency-domain analysis (in Matlab) on the measurement signal to identify the spurious components and in which frequency bands they lie. b) identify and motivate a suitable approach to remove the spurious components using digital filters; the design of the filters can be based on methods discussed in detail in the course as well as on further design approaches available in the literature. c) based on (b), design the relevant filters (i.e use Matlab to specify appropriate filter coefficients) and apply these to the measurement signal (using the filter command in Matlab) in order to obtain a cleaner, noise-reduced signal that reveals characteristics similar to those displayed in Fig 1. d) by visual inspection of the resulting noise-reduced signal (you will need to zoom in for this) estimate as accurate as possible: the inter-peak interval (in milliseconds, rounded to an integer) the height of both the primary-peak and of the secondary-peak (in Volts, rounded to two decimal places) . The Measurement Signal. Each student is supplied with an individual, separate signal (i.e. none of the signals have the same parameters). The signals are available in *.wav format on Canvas, listed as a table on the page of the Coursework Assignment 2. Please download the *.wav signal assigned to you per student number. You can import the signal in matlab using the command (substitute the relevant number for X): [y,Fs] = audioread(‘signalX.wav’); Where y is the vector containing the signal (in Volts) and Fs is the sampling frequency (which is set at 2000 Hz). Note that the presence of spurious signal components means that the measurement signal will, in raw unfiltered form, mostly resemble noise when plotted. Assignment Submission The submission (electronically, on Canvas) is a short report in PDF format (maximum 10 pages including appendix) which documents your work on the all of the above subtasks. The report should: use 11-point Arial font (or similar) and single-line line-spacing, and the text width should be close to normal standards. have the assignment title and your name and student number on the front page. feature result plots generated in Matlab to support your explanations and demonstrate your findings; these should be labelled as figures with a descriptive caption, and referred to in the report text; each figure should be legible (e.g. appropriate fontsize and line width, and zoomed in to the relevant details where appropriate) and have axis labels and units; see Fig 1 for an example. The report should include plots of (i) the amplitude spectrum of the measurement signal, (ii) the amplitude responses of the designed filters, and (iii) several periods of the noise- reduced signal. have a structure in which sections align with the four subtasks listed above under “Assignment Brief”. The report also needs a (brief, single-paragraph) introduction outlining the problem and a (brief, single-paragraph) conclusion section summarising the findings. An abstract is not required, but if you employ methods not already explored in detail in the course, you should refer the relevant literature listed in a Reference section. have an appendix that shows all your Matlab code for the assignment, which is preferably in a single Matlab script; use appropriate commenting to make your code more readable. Where appropriate, you should make use of in-built Matlab functions: many of the potentially useful in-built functions are listed on the final page of session Lecture notes. Where appropriate, the main body of the report should make reference to the Matlab code or part of that. Assessment Criteria The report will be assessed according to the following criteria and weights: Correctness and interpretation of the results 30% Choice and motivation of the methodology 20% Correctness and readability of the Matlab code 20% Clarity, writing, organisation and formatting of the report 30% Assignment Mark The assignment will be marked out of 100%. The assignment counts for 20% of the total mark for ELE8059. Advice on Plotting in Matlab For incorporation of plots into MS Word, the following Matlab commands are useful: set(gcf,’PaperPosition’,[10 10 26 10]); print -djpeg -r300 test.jpg This creates a jpeg file named “test.jpg” containing the current figure at reasonable resolution and setting a frame size of 26 by 10 (adjust these according to your needs). See here for further details on setting the frame size: https://uk.mathworks.com/help/matlab/creating_plots/save-figure-at-specific-size-and- resolution.html To plot a line of defined width, you can use plot(x,y,’LineWidth’,1.2); where 1.2 is now the line width (adjust this according to your requirements). Where needed, you can use various freely available software to crop the figure.
