Development

Weekly Progress

Week 1 (22/06-28/06)

  • Decide on logistics needed and purchasing them, taking size into consideration
    • Flexi PCB
    • Arduino MKR1000
    • Amplifiers
    • Microphone of 9.7 mm diameter
    • Battery
  • Draft schematics of our prototype and of the circuit
  • Found heart sound recordings to be used for the project

Week 2 (29/06-03/07)

  • Development of Arduino code for analysing sounds
  • Design of echo chamber on Fusion360
  • Sent in orders for materials and claimant forms

Week 3 (06/07-10/07)

  • Progress meeting 1
  • Design of circuit diagram
  • Finalising echo chamber to be sent for printing
  • Research on ideal shape and size for echo chamber
    • Ratio of diameter to height
    • Texture of inner walls

Week 4 (13/07-17/07)

  • Faced contact issues with the prototype due to incorrect assembly of circuit and parts
  • Resolved issues
  • Conducted tests on noise floors for circuit
    • No filter
    • 1 filter
    • 2 filters
  • Sent in files for 3D printing of echo chamber

Week 5 (20/07-24/07)

  • Decided on TPU to replace silicon due to cost factors
  • Faced issues with 3D printed echo chamber due to inexperience in 3D modelling and small size
  • Rewire code for Arduino in assembly code
  • Faced complications with audio encoding and recording, lack of samples for lung sounds

Week 6 (27/07-31/07)

  • Visited 3D printing vendor to enquire on issues faced and how to improve our models
  • Managed to find and bought diaphragm for echo chamber
  • Sent new files in for 3D printing
  • Started on development of mobile app to control wireless stethoscope using MIT App Inventor
  • Completed coding hardware
  • Implemented networking code
  • Found 2 FYPs with useful code for signal processing
    • Managed to filter lung sounds to detect only unhealthy lung sounds

Week 0 (03/08-07/08)

  • Progress meeting 2

Week 1 (10/08-14/08)

  • Debugged Arduino library
  • Lung sound processing
  • Soldered SD card and microphone to Arduino for testing

Week 2 (17/08-21/08)

  • Focus only on ‘normal’, ‘crackle’, ‘wheeze’ and ‘both crackle and wheeze’ for lungs sounds instead of looking at entire data set. Other lung sound scenarios (eg asthma) have too small sample size for accurate model training.
  • Tested out connection of microphone to computer to record heart / lung sounds
  • Tested out different echo chamber shapes with large diameters
  • Decided on horn and cylinder shapes

Week 3 (24/08-28/08)

  • Purchased silicon tape for microphone box
  • 3D designed small horns and cylinders for testing
  • Designed App to connect mobile device to stethoscope on MIT app inventor

Week 4 (31/08-04/09)

  • Tested small echo chambers and decided on the horn with shorter height (keeping the original height to diameter ratio)

Week 5 (07/09-11/09)

  • Code algorithm that can predict heart and lung sound statuses
  • Heart sound algorithm accuracy ~94%; Lung sound processing algorithm accuracy ~41%

Week 6 (14/09-18/09)

  • Designed box for prototype
  • PCB design

Week 7 (21/09-25/09)

  • Decided on new shape for box (curved rather than rectangular)
  • How to organise components inside box
  • 3D designed and sent new box for printing

Recess Week (28/09-02/10)

  • Progress meeting 3