Maximum likelihood joint angle and delay estimation in unknown noise fields

We address the problem of joint angle and delay estimation using a sensor array in an unknown additive noise field. We propose a stochastic maximum likelihood (ML) estimator. The algorithm which is a 2D extension of the approximate maximum likelihood (AML) is applied to the multiple channel sample model and exploits the shift invariance of the data. The model allows the estimation of more parameter pairs than sensors and robustness of the algorithm makes it possible to use blind techniques to estimate the channel. Basic performance of the ML estimator are assessed through simulations and are compared with other high resolution methods. Comparisons are made against the stochastic Cramer-Rao bound (CRB) which is derived in the appendix.