OpenPDR

An opensource repository of Pedestrian Dead Reckoning which is device-independent and sensor-independent.

Object

1. Standard Platform support for novel PDR algorithm development,verification and comparation

2. Dataset support for research on full scenes of pedestrian movement

3. Data Prepocessing Tools specified for PDR applications

3.1 Detection and removal of outliers

3.2 Sensor automatic calibration method

Ellipsoid fit method(6 parameters)
Six-sided Calibration
Allan variance analysis

3.3 Smoothing and filtering

Moving average filters(unwighted,wighted)
Median filters(unwighted,wighted)
Complementary filters
Kalman filters(KF/EKF/UKF)
Particle Filters(PF)

4. Typical Implementation of PDR algorithms derived from mainstream academic research

5. Interface open to integrate with other localization modules