アウトオブシーケンス測定を用いたトラッキングのためのアンサンブルカルマンフィルタ

概要

論文の詳細を見る
In target tracking problem, we can use multiple sensor to improve accuracy of the system. However, measurements from various sources are sometimes delayed and come at different rates. These measurements are known as out of sequence measurements (OOSMs). In other researches, Extended Kalman Filter (EKF), Particle Filter (PF) and Ensemble Kalman Filter (EnKF) provide us good solutions to OOSM problems. A solution via EKF is fast but cannot be applied to strong nonlinear systems. On the other hand, although PF is applicable to nonlinear system, the computational cost is so high that it is difficult to apply to real applications. In this paper, we propose techniques to solve OOSM problem using EnKF whose properties lay between EKF and PF. It is applicable to strong nonlinear system, requires less computation than PF. At the end of this paper, we investigate the performance of the solution via EnKF comparing to others. The results show competitive performance and attractiveness of solving OOSM via EnKF.