Maneuver target tracking with acceleration estimation using target past positions

Abstract

In air traffic control systems, the /spl alpha/-/spl beta/ filter and the Kalman filter have been employed for real-time digital tracking. It is desirable in multiple-target tracking to have small calculation burdens. Non-maneuvering targets can be accurately tracked with a constant velocity filter, however maneuvering targets might be missed due to the large tracking errors. For maneuvering target tracking, some acceleration estimation method such as the interacting multiple model (IMM) tracking and the two-stage Kalman filter have been proposed. However, their calculation requirements might be large. In this paper, we propose the /spl alpha/-/spl beta/ filter for a maneuvering target using acceleration estimation from the target past positions in order to obtain both good maneuver following capability and small computational requirements. We use simple target dynamic models for computer simulation, and show the effectiveness of the proposed filter in comparison with conventional methods. We also evaluate the quality of the proposed acceleration estimation method.