Track-level registration for networked trackers

Abstract

The paper presents a recursive algorithm for joint registration and track-to-track fusion based on equivalent measurements generated by geographically separated multitarget radar trackers. The input data for the algorithm are clutter-free decorrelated equivalent measurements and associated covariances that have been extracted from sensor-level track estimates. Simulation results show that the proposed algorithm adequately estimates sensor biases, and the resulting central-level track estimates are free of registration errors. Furthermore, equivalent measurements generated for this algorithm are also suitable for processing by existing batch-processing registration algorithms.