Estimating biases in sensor measurements using airlane information

Abstract

When sensors are poorly registered, systematic errors or biases can appear in their measurements, hampering the formation of a fused surveillance picture. To estimate and correct for these biases, a method exploiting airlane information is proposed. Models of the bias state and bias measurement are first formulated. Then, based on the airlane associated with a target of opportunity, a Gaussian mixture model is formulated for the target's position. Particle filter estimation is employed to handle the nonlinear/non-Gaussian nature of the models. Simulation results are given to demonstrate the ability of this method to correct for biases in sensor measurements effectively.