An explanation to the limitation observed in the Kumaresan-Prony algorithm
D. B. Rao · IEEE Transactions on Acoustics Speech and Signal Processing · 1986
In this correspondence we explain the limitation of the Kumaresan-Prony algorithm observed in [2]. It is shown that the condition number of the matrix involved with the associated least-squares problem periodically degrades (condition number \rightarrow \infin) making the least-squares solution ill conditioned. This ill conditioning causes a degradation in the estimates obtained by the Kumaresan-Prony algorithm.