Characterization of Cubature Kalman Filter for GPS Delayed Environments in INS-GPS Integrated Navigation

Abstract

Position estimation of an Unmanned Aerial Vehicle (UAV) can be done using inertial navigation system (INS). Even though there exists a number of sensors for inertial navigation, MEMS based sensors are preferred over others in applications related to UAVs because of their compact size and low cost. Inertial sensors alone can be used for navigation but as MEMS sensors causes error to accumulate over time, INS can be aided with GPS signal to improve its accuracy over time. Since GPS signals may not be available always, a better fusion algorithm which fuses the INS measurements with GPS signal is needed. In this paper, a fusion algorithm called Cubature Kalman filter (CKF) is used for the position estimation of the vehicle which is more reliable than the conventionally used Extended Kalman filter (EKF) algorithm. Both the algorithms are used to estimate the position of the vehicle and they are compared. It is evident from the analysis that CKF effectively improves the error in estimation compared to EKF.