/MCI_PA1

Primary LanguageJupyter Notebook

MCI_PA1

Stage 1 & 2

The android app code is in MCI_PA1_Stage1_2 folder. - The output CSV files are in result folder.

Stage 3

In MCI_PA1_Stage3 folder Step_counter.ipynb: Source code + Graph for Jupyter notebook version Step_counter.py: Source code for python version Step_counter.ipynb collaboratory.pdf: Jupyter notebook printed version

Basic Step counter Algorithm

  1. Gathering Accelerometer X,Y,Z data
  2. Applying a low-pass filter with Sample frequency 2Hz to eliminate random noises (2Hz = 12.5663706 Rad/s)
  3. Finding peak with three parameters; Height, prominence, and distance - Height: Minimal magnitude to count it as step. We can eliminate the small magnitude noises. - Prominence: Minimal magnitude difference with left and right valleys. Values depends on situatlion. - Distance: Minimal distacne from peak to peak. We will not count the steps too frequently. (Normaly 100 which means 200ms = 5Hz. I assume that human cannot walk faster than this)

Stage 4

The python code and output graphs are in MCI_PA1_Stage4 folder