LINEAR MINIMUM VARIANCE FILTER FOR PASSIVE LOCALIZATION

A simple and good post-processing method for passive localization is proposed. Post-processor of passive localization concerns estimating the time varying state of a dynamic system. The general way to do this is Kalman filter. This paper applies the Linear Minimum Variance （LMV） method, which is generally used for parameter estimation, to estimating the varying state of a linear dynamic system. So the new method can be called LMV filter. In fact, it is a weighted average method, because LMV filter takes weighted average of K samples of observation with different weight coefficients which are given by a system of equations. A basic assumption concerning source motion is to consider it to be of constant velocity with random perturbations over the observation intervals. Two simulation results are presenced to show: range estimate converges fast and good, no divergence appears, and the method has capacity to adapt to target maneuvre. Another important feature is its very low computational level, which is useful in poor computer facility case.