Hybrid Extended Particle Filter (HEPF) for Integrated Civilian Navigation System

P. Aggarwal, Z. Syed and Dr N.El-Sheimy

Abstract:	system (INS) and Global positioning system (GPS), improves navigation parameters accuracy. Currently, integrated navigation systems are commonly implemented using Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF). The EKF assumes linear process and measurement models while UKF generates sigma points using the real mean and standard deviation of data. However, both EKF and UKF assume the noise to be Gaussian, which is unrealistic for highly nonlinear systems. To overcome these limitations, Particle Filter (PF) was proposed lately which is a nonparametric filter and hence can easily deal with non-linearity and non-Gaussian noises. In this paper, Hybrid Extended Particle Filter (HEPF) is developed as an alternative to the EKF to achieve better navigation accuracy for low-cost Micro Electro Mechanical Systems (MEMS) sensors. Experimental GPS/INS datasets consisting of GPS carrier phase data and inertial measurements from low-cost MEMS-grade Inertial Measurement Unit (IMU) is used to evaluate the proposed HEPF. The HEPF performance is compared to that of other estimation techniques such as the EKF. The results show that both HEPF and EKF provide comparable navigation results during periods without GPS outages. However in cases when GPS outages are simulated, HEPF performs much better than the EKF, especially when simulated outages are located during periods with high vehicle dynamics.
Published in:	Proceedings of IEEE/ION PLANS 2008 May 6 - 8, 2008 Hyatt Regency Hotel Monterey, CA
Pages:	984 - 992
Cite this article:	Aggarwal, P., Syed, Z., N.El-Sheimy, Dr, "Hybrid Extended Particle Filter (HEPF) for Integrated Civilian Navigation System," Proceedings of IEEE/ION PLANS 2008, Monterey, CA, May 2008, pp. 984-992. https://doi.org/10.1109/PLANS.2008.4570072
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In