Sub-Optimality of the Prefilter in Deeply Coupled Integration of GPS and INS

Abstract

A robust integrated receiver that integrates an Inertial Navigation System and a Global Positioning System and uses a Kalman filter to close the phase lock loops has been proposed recently in the form of an Ultra Tightly Coupled (UTC) receiver. One of the complexities of this receiver is establishing the mathematical relationship between the INS states and the GPS measurements. Many forms of the ultra tight coupling method have been proposed by linearizing these relationships to realize these highly non‐linear relationships in the integrated system. Here, we analyze the optimality of such linearization and quantize how much accuracy is lost in the prefilter of a UTC receiver. The prefilter that models the relationship between the GPS signal quantities and receiver measurements is presented. A Posterior Cramer Rao Bound (PCRB) is derived. Simulation results show that the variance of the estimates from the prefilter is higher than the PCRB by as much as an order of magnitude.

A Simplified Baseband Prefilter Model with Adaptive Kalman Filter for Ultra-Tight COMPASS/INS Integration

“…Expanding Figure 2 for a single channel, the architecture of federated ultra-tight GPS/INS integration is shown in the Figure 3 [7,11]. Prefilter and integrated navigation filter are two kernel components of this architecture.…”

“…In federated architecture, the large integrated Kalman filter is decomposed into two filters operating at different rates [ 7 ], as shown in Figure 2 [ 13 ]. The code and carrier tracking loops are completed in a baseband signal pre-processing filter (prefilter for short) in each channel, and the integrated navigation filter (master filter) is used to process the output of the prefilters and restrict the INS errors.…”

Sample-Wise Aiding in GPS/INS Ultra-Tight Integration for High-Dynamic, High-Precision Tracking

“…Babu, R. & Wang, J. concerned the quality of IMU-derived Doppler estimates for integration, and derived the relationship between the baseband I (in-phase)/Q (quadrature) and the INS error for performance analysis [ 20 ]. Several studies addressed the algorithms of the pre-filter and the centralized navigation filter [ 21 , 22 ]. Hwang, D.B., et al ., provided a unified frame applicable to several different algorithms of GPS/INS deep fusion [ 23 ].…”

A double-filter-structure based COMPASS/INS deep integrated navigation system implementation and tracking performance evaluation

“…As with an independent GPS receiver, an independent COMPASS receiver will face the challenge in choosing a bandwidth to satisfy both anti-jamming capability and dynamics adaptation. GPS/INS deep integrated navigation system was proposed to solve this problem [3][4][5][6][7][8][9][10]. Compared with a traditional loosely or tightly integrated GPS/INS integration, the deep GPS/INS integration fuses GPS baseband signals and INS information in a deeply coupled mode to guarantee GPS receiver's anti-jamming capability and dynamic adaptation simultaneously.…”

“…Compared with a traditional loosely or tightly integrated GPS/INS integration, the deep GPS/INS integration fuses GPS baseband signals and INS information in a deeply coupled mode to guarantee GPS receiver's anti-jamming capability and dynamic adaptation simultaneously. For information fusion, the GPS/INS or any other GNSS/INS deep integration can be divided into central architecture [3,4] and federated architecture [5,6]. The federated systems can be divided into coherent and non-coherent based on whether the baseband signals pass throu- gh code and carrier discriminators or not [7].…”