MAP recovery of polynomial splines from compressive samples and its application to vehicular signals

Akira Hirabayashi, Satoshi Makido, Laurent Condat, Dimitri Van De Ville, Vivek K. Goyal, Manos Papadakis
2013 Wavelets and Sparsity XV  
We propose a stable reconstruction method for polynomial splines from compressive samples based on the maximum a posteriori (MAP) estimation. The polynomial splines are one of the most powerful tools for modeling signals in real applications. Since such signals are not band-limited, the classical sampling theorem cannot be applied to them. However, splines can be regarded as signals with finite rate of innovation and therefore be perfectly reconstructed from noiseless samples acquired at,
more » ... imately, the rate of innovation. In noisy case, the conventional approach exploits Cadzow denoising. Our approach based on the MAP estimation reconstructs the signals more stably than not only the conventional approach but also a maximum likelihood estimation. We show the effectiveness of the proposed method by applying it to compressive sampling of vehicular signals.
doi:10.1117/12.2024039 fatcat:6lhhkas7hjckje4gs7lblqbegi