EXPERIMENTAL GPS/INS/PSEUDOLITE SYSTEM FOR KINEMATIC POSITIONING

Abstract

The requirements for accuracy and reliability of kinematic positioning have dramatically increased in the past few years. Stand-alone GPS can, in general, provide high accuracy and stability over time, however, may still fail in adverse conditions (jamming, high multipath, high ionospheric activity, etc.) or in the environments with obstructions or locations with weak satellite constellation (high altitudes). One example of rather demanding positioning accuracy requirements is in mobile mapping technology that has significantly expanded its market during the last decade of the twentieth century. In order to meet the stringent accuracy requirements and to provide a continuous trajectory, the most common means of positioning used in mobile mapping is an integrated GPS/INS (inertial navigation system) system, used to provide positioning and attitude information for land-based or airborne imaging sensors. GPS contributes high accuracy and long-term stability (under no losses of lock and properly resolved ambiguities), providing means of error estimation of the INS sensors. A GPS-calibrated inertial navigation system provides reliable bridging during GPS outages and supports the ambiguity resolution after the GPS lock is reestablished. Such a system works well if GPS gaps are not too frequent and not excessively long. However, in case of navigation in urban canyons or indoor mapping, there is usually very limited or no GPS signal. Consequently, a pseudolite (PL) array may be used to supplement the satellite signal.In this paper, an experimental GSP/INS/PL system is presented, with a special emphasis on the error spectrum and the navigation performance analysis, based on a medium-accuracy and high-reliability INS, limited GPS constellation, and varying number of pseudolites. Double difference GPS and PL phase observations are formed and used as a measurement update in the extended Kalman filter, facilitating tight integration of the navigation sensors. The simulated data and the actual test results in typical noisy environments are presented, including the geometry and the vehicle dynamics analysis, and their impact on the navigation accuracy.