17-03-2014, 08:13 PM
Abstract
The Bearings-Only Tracking problem (BOT) has
been studied for many years and UKF has extensively
been applied in passive non-maneuvering target
tracking. When target takes maneuver, the estimation
will diverge. STF can track the abrupt change of state
by tuning the gain matrix online. Based on the
advantages of STF and UKF, a new method, STF_UKF,
is proposed to track maneuvering target with
bearings-only measurements from two separate
stationary observers. Simulation results show that
STF_UKF is feasible and effective in tracking
maneuvering target, especially when initial state
deviation and measurement noise are great.
1. Introduction
The Bearings-Only Tracking problem (BOT) is a
topic of great interest with many applications in passive
surveillance systems. For the case of tracking
non-maneuvering target, this problem has been
satisfactory solved by batch processing (e.g. least
squares or maximum likelihood estimation). Recently
the recursive solutions, such as unscented Kalman filter
(UKF) have drawn great interest. UKF is promising in
nonlinear filtering problem and a variety of UKF-type
passive tracking filters [1-4] have been reported.
However, so far few are concerned with passive
maneuvering target tracking problem by using
bearings-only measurements.
In this paper is considered the passive
maneuvering target tracking problem with
bearings-only measurements from two separate
stationary observers. The ability of UKF in tracking
target with constant velocity has been studied
extensively. When the target carries out a maneuver,
the solution (estimation) will diverge. Strong tracking
filter (STF) [5] is robust to the model uncertainty and
can track abrupt change of state. Based on the
advantages of UKF and STF, a new method for
tracking maneuvering target with bearings-only
measurements is proposed in this paper.
The Bearings-Only Tracking problem (BOT) has
been studied for many years and UKF has extensively
been applied in passive non-maneuvering target
tracking. When target takes maneuver, the estimation
will diverge. STF can track the abrupt change of state
by tuning the gain matrix online. Based on the
advantages of STF and UKF, a new method, STF_UKF,
is proposed to track maneuvering target with
bearings-only measurements from two separate
stationary observers. Simulation results show that
STF_UKF is feasible and effective in tracking
maneuvering target, especially when initial state
deviation and measurement noise are great.
1. Introduction
The Bearings-Only Tracking problem (BOT) is a
topic of great interest with many applications in passive
surveillance systems. For the case of tracking
non-maneuvering target, this problem has been
satisfactory solved by batch processing (e.g. least
squares or maximum likelihood estimation). Recently
the recursive solutions, such as unscented Kalman filter
(UKF) have drawn great interest. UKF is promising in
nonlinear filtering problem and a variety of UKF-type
passive tracking filters [1-4] have been reported.
However, so far few are concerned with passive
maneuvering target tracking problem by using
bearings-only measurements.
In this paper is considered the passive
maneuvering target tracking problem with
bearings-only measurements from two separate
stationary observers. The ability of UKF in tracking
target with constant velocity has been studied
extensively. When the target carries out a maneuver,
the solution (estimation) will diverge. Strong tracking
filter (STF) [5] is robust to the model uncertainty and
can track abrupt change of state. Based on the
advantages of UKF and STF, a new method for
tracking maneuvering target with bearings-only
measurements is proposed in this paper.