Radar Based Navigation in Unknown Terrain.

摘要

There is a great need to develop non-GPS based methods for positioning and navigation in situations where GPS is not available. This research focuses on the development of an Ultra-Wideband Orthogonal Frequency Division Multiplexed (UWB-OFDM) radar as a navigation sensor in GPS-denied environments. A side-looking vehicle-fixed UWB-OFDM radar is mounted to a ground or aerial vehicle continously collecting data. A set of signal processing algorithms and methods are developed which use the raw radar data to aide in calculating the vehicle position and velocity via a simultaneous localization and mapping (SLAM) approach. The radar processing algorithms detect strong, persistent, and stationary reflectors embedded in the environment and extract range/Doppler measurements to them. If the radar is the only sensor available, the measurements are used to directly compute the vehicle position. If an existing navigation platform is available, the measurements are combined with the other sensors in an EKF. The developed algorithms are tested via both a series of airborne simulations and a ground- based experiment. The computed navigation solution performance is analyzed with the following sensor availability: radar-only, INS-only, and combined radar/INS. In both simulations and experimental scenarios, the integrated INS/UWB-OFDM system shows significant improvements over an INS-only navigation solution.