ENGP5300/ENGP5302 DSP and Embedded Systems/DSP, Control and Instrumentation Coursework Assignment 2026

ENGP5300/ENGP5302 DSP and Embedded Systems/DSP, Control and Instrumentation

Coursework Assignment

What to submit:

Your coursework must be submitted as a report in Word format. Your MATLAB codes must be provided as text (notpictures) and included as appendices at the end of the report. All codelines must be commented to explain them. Your answers must besupported by references as appropriate. All references must be in IEEE style. Please note that4 marks will be allocated to the presentation andorganisation of your report. Your report, excluding the cover page and MATLAB codes, must not exceed 12 pages (minimum font size 11 pt).

Part I

a) Write a MATLAB code that measures the CPU time needed to findthe nearest neighbour of every point in a 3D point cloud usingi) brute force and ii) the kd-tree method (seeKDTree Searcher -Create Kd-tree nearest neighbor searcher - MATLAB). The querypoints must be the points in the point cloud, and you must excludethe point itself from being returned as its own nearest neighbour.Do this comparison for the following point clouds: Stanford Bunny,Horse, Atenea, and Venus, which are available here:3D Models. Discuss your results.(5 marks)

b) Provide a three-minute video explaining and demonstrating yourwork. Your video should show your MATLAB workspace and codeexecution.(3 marks)

Part II

a) You are given a noisy audio signal (speech_noisy.wav). Design and implement a digital filter in MATLAB that aims at removing thenoise while preserving speech intelligibility. Begin by analysing the signal’s frequency content (use the magnitude spectrum andspectrogram). Then, justify your chosen filter type and cutofff requencies. Demonstrate the effect of your filter on the signal’sfrequency content and speech quality.(5 marks)

b) Provide a three-minute video explaining and demonstrating your filter. Your video should show your MATLAB workspace and codeexecution.(3 marks)

Part III

Using JPEG and JPEG2000, compress the MATLAB built-in image onions.png at four bitrates. For each bitrate, compute the PSNR andcompare visual artifacts. Justify your design and discuss the results ofyour experiment.(5 marks)