A homotopy continuation approach for self-calibration of arrays with general phase perturbations

Abstract

An eigenstructure approach is presented for array processing with imperfect sensors. Each sensor is modeled as having an unknown angularly dependent phase error with ideal gain. The proposed technique is capable of compensating for sensor location error, unknown sensor phase patterns, and various types of environmental inhomogeneities. The algorithm can be used in self-calibration applications since the sensor perturbations are estimated in addition to angle of arrival of the sources. Based on the method of homotopy continuation it will find all solutions for the minimum. Since the algorithm is computationally intensive, a sub-optimal less intensive approach is also provided.<>