On PNT Integrity in Snapshot And Recursive Positioning Algorithms for Maritime Applications

Abstract

Resilient provision of Position Navigation and Time (PNT) data is a strategic key element of the e-Navigation strategy, developed by the International Maritime Organization (IMO). The improvement and the indication of reliability have been identified as high level user need with respect to PNT data supplied by electronics means. IMO as developed a maritime PNT system concept aiming to improve the resilience and reliability of PNT data provision during berth-to-berth navigation. The maritime PNT System comprises several structural components, where Global Navigation Satellite Systems (GNSS), have become the primary component to produce position, velocity and time information for maritime applications. For a comprehensive onboard provision of PNT data as well as to compensate the vulnerability of GNSS, further onboard sensors are needed. The PNT system is responsible for the fusion of the data provided by all the available onboard sensors and data integrity monitoring functions. A unit composed by several sensors of different classes improves the resilience of the system. DLR has developed a prototype of an onboard PNT unit and several measurement campaigns have been performed. This paper concentrates on integrity monitoring (IM) for navigation systems based on sensor fusion. IM is a mechanism that protects the user from large position and velocity errors in the presence of failures or non-scheduled events in a timely fashion. It can be seen as an instantaneous decision criterion for using or not the system and therefore constitutes a key function for the safety of navigation. The IM includes the detection and exclusion functions, they are responsible for detecting the measurements errors (faults) and exclude them from the PNT data computation algorithm. This work presents a systematic analysis of Fault Detection and Exclusion (FDE) algorithms in representative single and multi-sensor. More specifically, a pure GNSS-based snapshot weighted iterative least-square (WLS) solution is compared to a classical error-state Extended Kalman Filter (EKF) for a combined GNSS/IMU system with Euler angles for attitude parameterization. The outlier detection functionality is implemented for both pseudorange and Doppler shift observations in order to ensure the integrity of the estimated position and velocity data. The work confirms the superiority of the recursive Bayesian scheme over a snapshot algorithm in terms of the outlier detection performance. This can be explained by the recursive structure of the estimator, where the dynamical model of the system provides the additional source of information, which increases the system’s redundancy and hence improves the performance of the FDE schemes.