Unscented Kalman Filter with Application to Bearings- Only Target Tracking

Abstract

AbstractThe unscented transformation coupled with certain parts of the classic Kalman Alter, provides a more accurate method than the Extended Kalman Filter for nonlinear state estimation. Using bearings-only measurements, the unscented Kalman Filter algorithm estimates target motion parameters and detects target maneuver, using zero mean chi-square distributed random sequence residuals, in a sliding window format. During target maneuvering, the co-variance of the process noise is sufficiently increased in such a way that the disturbance in the solution is minimized. When target maneuver is completed, the covariance of process noise is lowered. The performance of this algorithm is evaluated using Monte Carlo simulation and results are presented.