期刊
IEEE SENSORS LETTERS
卷 7, 期 11, 页码 -出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/LSENS.2023.3326132
关键词
Sensor signal processing; bearing only tracking (BOT); Gaussian filters; observer maneuver; target motion analysis (TMA)
This letter proposes an ownship maneuver to improve the accuracy of state estimation when tracking a target with bearing only measurements. By maximizing the Fisher information matrix, the recommended maneuver leads to a reduced Cramer-Rao lower bound (CRLB) and improved estimator accuracy. It also considers certain heuristics to maintain safety and prevent the target from entering blind zones.
It is widely acknowledged that when tracking a target with nearly constant velocity using bearing only measurements, the observer must execute a maneuver to achieve observability and estimate states accurately. In such a scenario, the precision of state estimation algorithms is notably influenced by the trajectory the ownship takes during maneuvering. This letter aims to enhance estimation accuracy by recommending an ownship maneuver. The recommended observer maneuver, determined by maximizing the Fisher information matrix, leads to a reduced Cramer-Rao lower bound (CRLB) and improved estimator accuracy, evaluated in terms of root mean square error and track loss percentage. Certain heuristics are taken into consideration to maintain a not more than two-legged observer trajectory, to prevent the target from entering the blind zone of the sonar, and to avoid a direct collision of the observer with the target. Thus, this letter recommends a safe maneuver for the ownship to track an enemy target more efficiently. As the Fisher information matrix requires the true states to be known, this work has been done using offline true data of the target. The recommended maneuver for every course of the target can be saved in an onboard computer, and during real life engagement, the ownship can follow such maneuver recommendation once the course of the target is estimated.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据