A General Approach to Online Sensor Calibration for Navigation Using Factor Graphs

Abstract

We present a general method for online sensor calibration using factor graphs, which can be applied to a wide range of sensors and parameter types. Our approach refines a particular sensor’s calibration parameters using all sensors’ measurements recorded in a small period of time at the beginning of a navigation mission, and then uses the refined parameters in the rest of the mission. The parameter estimation module is implemented in our twotier navigation system which comprises a low-latency Hierarchical Multi-Rate Filter (HMF) and a high-quality Trajectory Smoother/Mapper (TSM). HMF and TSM run in parallel to provide a low-latency and accurate navigation solution. TSM includes the parameter estimation sub-module, which refines the sensor calibration parameters. We present methods for calibrating four types of parameters that are often encountered in a navigation system: wind velocity, compass bias, time offset between system’s clock and camera’s clock, and step size of pedestrians. In addition, we validated our proposed calibration method using both real aerial and ground data.