Reconstruction of sparse Legendre and Gegenbauer expansions (Q329024)

The authors present a new deterministic approximate approach to the reconstruction of sparse Legendre expansion when relatively few samples on a special grid are given. Using asymptotic properties of Legendre polynomials, this reconstruction is based on Prony-like methods. The key idea is that the conveniently scaled Legendre polynomials behave similar to the cosine functions near zero. Thus a local approximation of Legendre polynomials by cosine functions is possible using a sampling grid located near zero. Numerical examples are given including one showing that the method proposed is robust with respect to noisy sampled data. Finally, they extend the suggested method to the reconstruction of sparse Gegenbauer expansions of low positive order.