Non-asymptotic statistical behaviours of ´ML´ methods for bearing estimation

The authors analyze the Cramer Rao bound and the statistical performance of some methods of direction-of-arrival (DOA) estimation, which can be derived from the maximum likelihood method (MLM). They use only ensemble average on time for all definitions concerning the second-order properties of the amplitudes. The result is that the further derivations are valid in the small sample case, even for one snapshot. Statistical behaviour of ML-based methods is established conditionally to a particular realization of the signal emitted. For this, a theoretical variance is derived of ML-based methods which allows an easy study of parameter influence on performance.<>