One basic requirement of autonomous operation in distributed satellite system (DSS) is autonomous navigation systems which is to determine the position and velocity based solely on the sensors onboard. Several types of measurements have been used in autonomous navigation systems such as relative position vector, inter-satellite range and bearings. A general method of autonomous navigation using relative position is the extended Kalman filter (EKF) algorithm. One problem of the EKF algorithm is the computational load because usually all satellites’ position and velocity are included in the state vector which can be really large. To solve this problem, this paper investigates a two-stage extended Kalman filter (TSEKF) algorithm which decouples the augmented filter into two parallel lower-order filters. The number of floating-point operations, i.e. FLOPS, is calculated to compare the computational load of TSEKF to that of general EKF algorithm. From the simulation result, we can get that the TSEKF has an equivalent performance in navigation accuracy with a much less computational complexity.
Paper
Full text
A Two-stage Extended Kalman Filter for Autonomous Navigation of Two-satellite System
Semantic Scholar · Engineering · 2020
Abstract
One basic requirement of autonomous operation in distributed satellite system (DSS) is autonomous navigation systems which is to determine the position and velocity based solely on the sensors onboard. Several types of measurements have been used in autonomous navigation systems such as relative position vector, inter-satellite range and bearings. A general method of autonomous navigation using relative position is the extended Kalman filter (EKF) algorithm. One problem of the EKF algorithm is the computational load because usually all satellites’ position and velocity are included in the state vector which can be really large. To solve this problem, this paper investigates a two-stage extended Kalman filter (TSEKF) algorithm which decouples the augmented filter into two parallel lower-order filters. The number of floating-point operations, i.e. FLOPS, is calculated to compare the computational load of TSEKF to that of general EKF algorithm. From the simulation result, we can get that the TSEKF has an equivalent performance in navigation accuracy with a much less computational complexity.