stanford graduate
course project

AA 272 Global Positioning Systems

winter 2022

In a graduate course, Global Positioning Systems, taught by Dr. Grace Gao, we studied the principles of satellite navigation. We learned how raw signals are processed and analyzed to produce a position solution

For the final project, my group combined inertial and GPS measurements to estimate a pedestrian's position. Measurements were combined using a complementary filter and an extended Kalman filter.

my work

Our test apparatus consisted of an off-the-shelf IMU with an Arduino Uno.

Characterization of the IMU in a nominal, stationary configuration is required. The corresponding Allen variance plots are included from the IMU characterization tests.

This illustration describes how the complementary filter uses a weighted average of measurements to predict a future state from a previous state.

The framework for applying an extended Kalman filter was a bit more complicated, but includes the transforming the IMU measurement to a global frame and then applying state estimation methods.

A data collection test of a pedestrian circling a field is shown below. The filtered methods outperform the predictions using GNSS measurements alone.

← →