Robust bearings-only tracking algorithm using structured total least squares-based Kalman filter

A nonlinear approach called the robust structured total least squares kalman filter (RSTLS-KF) algorithm is proposed for solving tracking inaccuracy caused by outliers in bearings-only multi-station passive tracking. In that regard, the robust extremal function is introduced to the weighted structured total least squares (WSTLS) location criterion, and then the improved Danish equivalent weight function is built on the basis, which can identify outliers automatically and reduce the weight of the polluted data. Finally, the observation equation is linearized according to the RSTLS location result with the structured total least norm (STLN) solution. Hence location and velocity of the target can be given by the Kalman filter. Simulation results show that tracking performance of the RSTLS-KF is comparable or better than that of conventional algorithms. Furthermore, when outliers appear, the RSTLS-KF is accurate and robust, whereas the conventional algorithms become distort seriously.