On an improved gradient technique for adaptive array processing

An improved adaptation algorithm designed for real-time data processing in antenna arrays in presented. The method is used for determining the set of filter weights in a sidelobe cancellation system. In order to solve the Wiener filter equation, the proposed method employs the gradient descent technique using a sample matrix estimate. An algorithm, the sample matrix gradient technique, therefore combines the convergence properties of the direct SMI-technique (to some extent) and the numerical benefits of an iteration method. Numerical stability and modest computational load of the iteration method make the proposed algorithm especially useful for large antenna array applications.