Skip to content

Low-Pass Filter

This class is designed to smooth out accelerometer data by removing high-frequency noise (quick and insignificant changes in movement). It uses a mathematical technique called a low-pass filter to focus on slower, meaningful movement patterns. This is useful when analyzing movement data because it makes trends and significant motions clearer.

Step-by-Step Explanation:

  1. Input Data:
    • The input is a table (DataFrame) with three columns: x, y, z. These represent movement measurements along three axes.
  2. Low-Pass Filter:
    • The filter removes fast, noisy signals from the data. For example:
      • If you’re walking, the filter keeps the steady up-and-down motion of your steps but removes tiny vibrations caused by your device or environment.
  3. How It Works:
    • A mathematical function (called a Butterworth filter) is applied to the data.
    • It keeps movement data below a certain speed (frequency), determined by the cutoff frequency. Movements faster than this speed are considered noise and are removed.
  4. Filter Settings:
    • Cutoff Frequency: The speed at which the filter decides whether to keep or remove a signal (e.g., 5 Hz means it keeps motions slower than 5 cycles per second).
    • Sampling Rate: The rate at which the device records movement data (e.g., 25 Hz means 25 data points per second).
    • Filter Order: Determines the sharpness of the filtering. A higher order makes the cutoff more precise.
  5. Smoothing the Data:
    • The class applies the low-pass filter to the x, y, and z columns of the data, smoothing them to remove noise.
  6. Output:
    • The filtered data is returned as a table (DataFrame) with the same structure as the input.
    • The smoothed columns (x, y, z) are ready for further analysis.
  7. Error Handling:
    • If the input data doesn’t include the required columns (x, y, z), the class raises an error to ensure proper usage.
  8. Friendly Messages:
    • After the filter is applied, the class prints a message: “Low-pass filter applied successfully,” so the user knows the process is complete.