Design an Optimized Nonlinear Integration Filter based on Particle Swarm Optimization algorithm for INS/GPS navigation system

Abstract: The Inertia Navigation Systems (INS) error, which today's basic navigation system for many uses, including military applications increases in time. Therefore, in order to achieve greater accuracy and reliability, especially in long-time navigation, such as in marine applications, an assistance system alongside the inertial navigation system should be used. In this case, the Global Positioning System (GPS) is the best navigational assistance system due to its complementary features. In an integrated navigation system consisting of a basic navigation system, along with a navigation assistance system, the Extended Kalman Filter (EKF) is a very common tool for integrating GPS and INS data. However, measurement and process noise covariance matrices are two important parameters in the Kalman filter, whose correct adjustment during navigation estimation process is very important in reducing the estimation error. In addition, since the governing equations of the inertial system are inherently nonlinear, the process of linearization in the Kalman filter adds an error due to linear approximation. Hence, researchers are looking for alternative algorithms for the Kalman filter, and so far a lot of research has been done. In this paper, an estimator based on particle optimization algorithm (PSO) is used as a substitute for Kalman filter based estimator. In this way, as soon as GPS observations are received, the estimation error of the navigation system based on the PSO algorithm is minimized to estimate the best output for the system. The simulation results and their application to several practical databases show that the proposed method significantly improved the accuracy of the estimation of navigational states by the integrated navigation system compared with conventional integration methods, such as the extended Kalman filter.