Sensor bias estimation from measurements of known trajectories

In this paper, we consider the problem of estimating sensor biases (e.g., range and bearing biases) from measurements of targets flying on known trajectories (i.e., zero, or near-zero, process noise). The key difficulty with this form of sensor registration is that the a priori kinematic uncertainty of the target state is often inaccurate. In this paper, we examine the sensitivity of two bias estimators: the extended Kalman filter (EKF) and nonlinear least squares (NLS) estimator, to the precision of the a priori information about the object. We examine a simplified two-dimensional problem in order to simplify the calculations required in the NLS iteration. In addition, the performance of the estimators is compared with the derived Cramer-Rao lower bound (CRLB).