Abstract
A new Cramér-Rao bound (CRB) is derived for projection/separation based joint estimation problems. The determination of bounds for the achievable performance of signal projection/separation methods, which specifically take into account the separated signal premise, is not an adequately addressed one, given the booming relevance of such methods in signal processing and wireless communication applications. In this reported work such a bound is derived, which is robust to variations in noise variances. This bound is evaluated for an important wireless communication application pertaining to parameter estimation in the uplink of OFDMA proposed for 4G communication. The features of the proposed CRB have been substantiated through computer simulation studies.