Extrapolation of discrete bandlimited signals in linear canonical transform domain

• Extrapolation algorithms for discrete (a,b,c,d)-bandlimited signals are proposed.
• Minimum energy least squares error extrapolation estimate are achieved.
• Effectiveness of the proposed extrapolation algorithms are verified.

This paper investigates the extrapolation problem of discrete bandlimited signals in linear canonical transform domains. First, by the generalized inverse theory of matrix, two extrapolation formulae which guarantee the minimum energy least squares error estimates are proposed. Then an iterative extrapolation algorithm is proposed, and by use of the interesting properties of discrete generalized prolate spheroidal sequences, its convergence to the minimum energy least squares error estimate is proved. Finally, numerical simulation results are given to demonstrate the effectiveness of the proposed extrapolation algorithms.