/sensor-fusion

Use accelerometer and gyroscope data from smartphones to identify vehicle type (bus or car) and phone location (driver side or passenger side).

Primary LanguagePythonMIT LicenseMIT

sensor-fusion

Use accelerometer and gyroscope data from smartphones to identify vehicle type (bus or car) and phone location (driver side or passenger side).

  1. Extract gravity signal (see "Extract Gravity Signal" jupyter notebook)
  2. Rotate XYZ signals to vehicle reference frame (see "Rotate Sensor Data to Vehicle Reference Frame" jupyter notebook)
  3. Resample time series data to 10 Hz sampling rate (see "Resample Sensor Data to 10 Hz Sampling Rate" jupyter notebook)
  4. Automate steps 1-3, modify columns to standard system, and save result to a new file (see "Process Smartphone Sensor Data" jupyter notebook)

Two examples of use cases of this software:

  1. Vehicle classification exercise: determine whether the smartphone is on a bus or in a car based on 5-10 minutes of sensor data (see "Vehicle Classification Exercise" jupyter notebook)
  2. Phone position classification exercise: determine whether the smartphone is on the driver side or passenger side of a car based on 5-10 minutes of sensor data (see "Phone Position Classification Exercise" jupyter notebook)