Iterated debiased Kalman filter for target tracking with converted measurements

Existing Kalman filters using unbiased converted measurement may still give biased estimates because the converted measurement covariance used is measurement noise dependent. An iterated converted measurement Kalman filter (ICMKF) is proposed to suppress this dependence, which at each recursion performs the measurement update twice. By reevaluating it at the estimated position at the second iteration, the covariance becomes much less noisy. Comparison with existing CMKFs demonstrated that the ICMKF has the best filtering performance.