Performance of GPS and INS integration algorithms under interference

Abstract

This paper carries out a comparative performance analysis of integration algorithms for navigation using global positioning systems (GPS) and inertial navigation systems (INS) when GPS is subject to intentional or unintentional interference. The effect of interference on GPS measurements is modelled as Gaussian additive noise, thus interference level is assumed to be low enough so that GPS is still operating, but high enough to degrade the quality of GPS. The paper investigates some of the existing centralised and decentralised GPS/INS integration algorithms such as the ones described in (Brown & Hwang 1997). The performance of these algorithms is analysed via Monte Carlo simulation studies using a Matlab/Simulink software package developed by the authors. Among different integration architectures, the simulation results indicate that the centralised integration algorithm, based on the complementary Kalman filtering, outperforms the decentralised ones when there is no interference. However, in the presence of GPS interference, one of the decentralised integration algorithms seems to be more robust than the others. During simulations, it has been observed that the model mismatch due to interference may cause filter divergence. This paper demonstrates that the GPS performance can be improved significantly by integrating it with an INS when it is subject to interference.