Navigation via H-field signature map correlation and INS integration

There is a great need to develop alternative sensors for vehicular navigation when GPS is not available. While the use of active on-board radars for SLAM and beacon navigation is an ongoing area of research, there are many situations where non-emitting sensors are desirable for navigation. This paper presents a highly available, passive-only navigation solution via longtime correlation of the static H-field against previously recorded magnetic field maps. Our solution is a two step process. In step one, we build a map of the magnetic field in the local area and calibrate the data to account for ferrous materials in the car and extrinsic alignment errors. In step two, we use the magnetic field map to position a vehicle in real-time without GPS, using passive-only sensors. Our solution is first developed theoretically and then validated experimentally in a road test. The road test is performed using a navigation platform consisting of a magnetometer, an INS, and a GPS receiver rigidly fixed to a ground vehicle. The vehicle is driven on a set of roads in the Athens, OH area, recording 3-axis magnetic field strengths, inertial data and GPS positions. This data is then processed in order to construct the magnetic field map in the local region. The car is then driven again on the same set of roads collecting measurements from the INS and the magnetometer. Using a particle filter, the INS and magnetometer measurements are combined with the magnetic field map to compute a position solution of the vehicle over time. The vehicle's navigation solution is then compared to the integrated truth solution using GPS to determine the error of the navigation solution. The results demonstrate the effectiveness of a robust, velocity-aided, static magnetic field/INS-based navigation solution.