Multiple invariance cumulant ESPRIT for DOA estimation

In this paper, cumulant based direction of arrival (DOA) estimation using multiple invariances is proposed which results in Multiple Invariance Cumulant ESPRIT (MICE) algorithm. In all previous formulations of cumulant based ESPRIT, only one invariance is exploited for DOA estimation. The cumulant matrix (if chosen properly) inherits the multiple invariance property if multiple displacement invariances are present in the sensor array. DOA estimation can be improved by exploiting these invariances simultaneously. A subspace fitting based fitness function is developed which simultaneously incorporates these multiple invariances. MICE depends on the effective minimization of this fitness function. Newton's method based minimization of this fitness function leads to the cumulant counterpart of (second order) Multiple Invariance ESPRIT algorithm (MI ESPRIT). Genetic Algorithm based minimization of this fitness function has also been investigated and shown to have various advantages. Simulation results are presented to show the effectiveness of the proposed method.