An Efficient Filter Bank Approach for the Provision of Integrity in Integrated Navigation Systems

摘要

Integrity information, i.e. upper bounds for the position error that are exceeded only with a certain probability, is required in a variety of applications. For a snapshot GNSS PVT, RAIM algorithms are widely used to solve this problem. In integrated navigation systems, in which different navigation sensors and techniques are combined, these RAIM algorithms cannot be used, because RAIM algorithms require the PVT to be obtained using snapshot least squares. In this paper, an efficient filter bank approach is proposed for the provision of integrity information in integrated navigation systems. By modifying a special type of filter bank, the Generalized Pseudo-Bayesian 1 (GPB1) filter bank, it is possible to perform just one single prediction step for all elemental filters of the filter bank. This leads to a significant reduction in the computational load compared to other filter bank approaches. The viability of the proposed approach is demonstrated in numerical simulations of an integrated navigation system combining a GNSS receiver, a z-gyro and an odometer.