An EM algorithm for multisensor TDOA/DD estimation in a multipath propagation environment

The problem of estimating time-difference-of-arrival (TDOA) and differential Doppler (DD) in a multipath environment is addressed. Correlated multipath interference has degenerative effects on the cross-ambiguity function which is the conventional joint TDOA/DD estimator. A new method, previously proposed for multisensor/multipath TDOA estimation, is extended to accommodate linearly-varying time delays. The approach is based on the expectation-maximization (EM) algorithm. Nonlinear optimization of the complete-data likelihood function uses the efficient RELAX method of estimating the frequencies, amplitudes, and phases of complex sinusoids in noise. The structure of the algorithm is described and its performance is demonstrated through computer simulation