Inertial Navigation System/Global Navigation Satellite System (INS/GNSS) integration system have been widely applied in recent years. Unfortunately, it sometimes malfunctions and the performance heavily deteriorates, especially in urban area where signals from satellites may be blocked or reflected by modern buildings. In multipath or Non Light-of-sight (NLOS) environment, incorrect signal results in poor observability of GNSS measurement model in Kalman Filter (KF). For purpose of addressing the issue, we proposed an adaptive strategy-based tightly-coupled INS/GNSS integration system aided by odometer and barometer, targeting to mitigate the error from poor observability. In this method, tightly-coupled (TC) scheme is implemented as the fundamental system in order to increase the reliability and stability. TC is more suitable than Loosely-coupled (LC), the traditional scheme, in urban navigation because it requires less visible GNSS measurement and it overcomes the disadvantage of LC, and further enhances the navigation result. Furthermore, aiding sensors such as odometer and barometer are integrated in this system as well, serving as velocity and height constraints respectively. Since the precision of GNSS positioning depends on the properties of the environment, measurement model of KF must work adaptively. Thus, innovation-based Adaptive Scaled Estimation (IASE) and Residual-based Adaptive Scaled Estimation (RASE), are also implemented to improve navigation performance in this paper. Finally, from the experimental validation, the proposed adaptive sensor-fusion navigation algorithm significantly enhanced the performance. The improvement was approximate 80% compared with the pure TC scheme; the RMSE can reach 6m in 3D and 2.5 m in vertical.
|頁（從 - 到）||881-888|
|期刊||International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives|
|出版狀態||Published - 2019 6月 4|
|事件||4th ISPRS Geospatial Week 2019 - Enschede, Netherlands|
持續時間: 2019 6月 10 → 2019 6月 14
All Science Journal Classification (ASJC) codes