Oversampled Adaptive Sensing via a Predefined Codebook

TL;DR

This paper extends the oversampled adaptive sensing (OAS) framework to realistic scenarios where sensing coefficients are chosen from a predefined codebook, demonstrating improved performance over classical non-adaptive methods.

Contribution

The paper introduces a new OAS variant that operates with a predefined codebook of sensing coefficients, broadening its practical applicability.

Findings

OAS with a codebook outperforms non-adaptive sensing methods.

The proposed method effectively utilizes prior information for sequential belief updates.

Performance gains are demonstrated in structured signal sensing scenarios.

Abstract

Oversampled adaptive sensing (OAS) is a Bayesian framework recently proposed for effective sensing of structured signals in a time-limited setting. In contrast to the conventional blind oversampling, OAS uses the prior information on the signal to construct posterior beliefs sequentially. These beliefs help in constructive oversampling which iteratively evolves through a sequence of time sub-frames. The initial studies of OAS consider the idealistic assumption of full control on sensing coefficients which is not feasible in many applications. In this work, we extend the initial investigations on OAS to more realistic settings in which the sensing coefficients are selected from a predefined set of possible choices, referred to as the codebook. We extend the OAS framework to these settings and compare its performance with classical non-adaptive approaches.