Systolic array architecture for adaptive eigenstructuredecomposition of correlation matrices

摘要

Eigenstructure decomposition of correlation matrices is animportant pre-processing stage in many modern signal processingapplications. In an unknown and possibly changing environment, adaptivealgorithms that are efficient and numerically stable as well as readilyimplementable in hardware for eigendecomposition are highly desirable.Most modern real-time signal processing applications involve processinglarge amounts of input data and require high throughput rates in orderto fulfil the needs of tracking and updating. The authors consider theuse of a novel systolic array architecture for the high throughputonline implementation of the adaptive simultaneous iteration method(SIM) algorithm for the estimation of the p largest eigenvalues andassociated eigenvectors of quasi-stationary or slowly varyingcorrelation matrices