2018/11/15 by Ralf R. Müller, Müller, Ralf R., Ali Bereyhi +3
Computer Science · Engineering · #Blind Source Separation Techniques #Distributed Sensor Networks and Detection Algorithms #FOS: Computer and information sciences #Information Theory (cs.IT) #Sparse and Compressive Sensing Techniques
paper · pdf · doi:10.48550/arxiv.1811.06472
openalex publication_date 2018/11/15 · openalex created_date 2022/08/02 · openalex updated_date 2026/07/28
Oversampled adaptive sensing (OAS) is a recently proposed Bayesian framework\nwhich sequentially adapts the sensing basis. In OAS, estimation quality is, in\neach step, measured by conditional mean squared errors (MSEs), and the basis\nfor the next sensing step is adapted accordingly. For given average sensing\ntime, OAS reduces the MSE compared to non-adaptive schemes, when the signal is\nsparse. This paper studies the asymptotic performance of Bayesian OAS, for\nunitarily invariant random projections. For sparse signals, it is shown that\nOAS with Bayesian recovery and hard adaptation significantly outperforms the\nminimum MSE bound for non-adaptive sensing. To address implementational\naspects, two computationally tractable algorithms are proposed, and their\nperformances are compared against the state-of-the-art non-adaptive algorithms\nvia numerical simulations. Investigations depict that these low-complexity OAS\nalgorithms, despite their suboptimality, outperform well-known non-adaptive\nschemes for sparse recovery, such as LASSO, with rather small oversampling\nfactors. This gain grows, as the compression rate increases.\n