Signal parameter estimation via the Cayley-Hamilton constraint

The estimation of signal parameters via rotational invariance technique (ESPRIT) algorithm has proved to be a robust solution to exploit single invariance structure in array elements. However, ESPRIT is not flexible to incorporate multiple invariances. A closed form solution to the problem of estimating eigenvalues in multiple invariant structures is presented in this paper. It is shown that the algorithm is comparable in complexity to ESPRIT but performs better than ESPRIT, especially when only a limited number of signal samples are available. In addition, the algorithm is shown to be feasible even for single array elements sharing multiple invariance.