|Robust UKF-IMM Filter for Tracking an Off-road Ground Target
Jae Weon Choi* and Kangwagye Samuel
International Journal of Control, Automation, and Systems, vol. 17, no. 5, pp.1149-1157, 2019
Abstract : "In this paper, the design of a robust UKF-IMM filter for tracking a fulltime off-road ground target is
studied and carried out. A detailed description of the target dynamic systems with respect to motion modelling
is done for a sharply maneuvering target. A four model IMM filter with two discrete white noise acceleration
models and two horizontal coordinated turn models is proposed. In this IMM filter, a low noise and high noise
model is proposed for the two pairs of motion models. Low noise model for non-maneuvering or slow maneuvering
situations and high noise model for quick and sharp maneuvering situations. The Unscented Kalman filter is used
as the base filter due to the highly nonlinear horizontal coordinated turn model with unknown turn rates along with
linear Kalman filter for linear systems. Simulation is carried out and results are presented showing the performance
of the proposed estimator and demonstrates its capability to perform its designed purpose. The simulated results
show that the four model IMM filter can track the highly maneuvering off-road target with acceptable error margin.
The error dynamics are observed to be stable with good maneuver detection characteristics. The proposed filter has
a low computation complexity and therefore can be adapted to most computers."
"Kalman filters, maneuvering target, off-road ground target, radar target tracking, robust IMM filter, state estimation, unscented Kalman filter"
Download PDF : Click this link