Dual-Constellation Aided High Integrity and High Accuracy Navigation Filter for Maritime Applications

Abstract

This publication presents the development of integrity monitoring and fault detection and exclusion (FDE) of pseudorange measurements, which are used within a tightly-coupled navigation filter especially designed for maritime applications. The developments are part of the research project GALILEOnautic 2 (grant number 50NA1808). The filter is based on an inertial measurement unit (IMU), is aided by a Doppler velocity log (DVL), and signals of Global Navigation Satellite System (GNSS). Particularly, the navigation filter processes GNSS observables of two systems: the Global Positioning System (GPS) and the new European satellite navigation system Galileo. GNSS observables are Doppler- and pseudoranges. To ensure high integrity and accuracy for the filter, it is crucial to provide a set of GNSS measurements for the navigation filter with sufficient observed satellites and observables without failure. Therefore, Weighted Receiver Autonomous Integrity Monitoring (WRAIM) is implemented, which is able to check the consistence of GNSS measurements and to identify and exclude faulty signals. Due to the synchronization of GPS and Galileo observables using the reliable Galileo GGTO message, an uncertainty is eliminated before integrity monitoring is performed. As a result, the WRAIM-variables can be reduced. To evaluate the GNSS-measurement quality, the implemented WRAIM-approach calculates a covariance matrix dynamically considering elevation angles of the received satellites and current GNSS settings, such as GPS stand-alone, dual frequency or differential correction. For the purpose of real application and filter stability, WRAIM is extended with a designed fault-exclusion functionality, in order to provide a stable and reliable subset of input satellite measurements for navigation filter. Firstly, an online test has been executed in an urban environment with multi-signal disturbances in Aachen, Germany. Here, WRAIM, together with the introduced extensions, identifies and excludes these multi-failures. Additionally, all the sensor data and GNSS correction data are recorded online and post-processed, such that the test scenario is reproduced with variations, e.g. without WRAIM or without fault exclusion extension, to verify the accuracy improvement by using the designed approach. The results show, the developed WRAIM approach identifies and excludes these multi-failures and is, therefore, able to calculate a more accurate navigation solution. Secondly, WRAIM is evaluated with (pre-processed) data from an extensive real-world measurement campaign, which has been carried out in the Harbor of Rostock, Germany. The results show that faulty GNSS measurements can be successfully detected and identified. After excluding the faulty signal, the 2D accuracy is improved by 54 % with respect to the reference solution from a real time kinematic (RTK) capable receiver.