LAUSR

In this article, we address the problem of estimating with high precision the time difference of arrival (TDOA) and frequency difference of arrival (FDOA) for interleaved pulse trains radiated from multiple sources. The pulse signal parameters of multiple sources are indistinguishable and as a result conventional de-interleaving techniques fail to separate the pulses emitted from each source. By noting that cross-correlating pulses from the same source at their true TDOA and FDOA produces consistent phases, a high-precision joint TDOA-FDOA estimator is proposed. Within the developed technique, a histogram-based clustering procedure is introduced to obtain initial TDOA and FDOA estimates. Then, an expectation maximization (EM)-based method is utilized to refine their estimates, where the unknown source labels of pulse signals are hidden variables and they are estimated together with TDOA and FDOA iteratively. Simulation results reveal the superior performance of the proposed algorithm compared to existing techniques, at the cost of moderate computational complexity increment.