Signal Power Estimation Via Vector and Matrix Approaches

Signal power estimation using an array of sensors is needed in many applications and can be done via both Vector Approaches (VA) and Matrix Approaches (MA). We consider a modified adaptive VA (MAVA) and a modified adaptive MA (MAMA) for signal power estimation and for achieving prescribed peak sidelobe levels as well as having some control over the main-beam shape. The formulations of both MAVA and MAMA are convex optimization problems, and thus globally optimal solutions can be determined efficiently. We will show that MAVA and MAMA are equivalent and that they are robust to small sample size problems, the presence of coherent interferences, and steering vector errors.