The restless multi-armed bandit formulation of the cognitive compressive sensing problem

Saeed Bagheri, Anna Scaglione

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

In this paper we introduce the Cognitive Compressive Sensing (CCS) problem, modeling a Cognitive Receiver (CR) that optimizes the K projections of a N > K dimensional vector dynamically, by optimizing the objective of correctly detecting the maximum number of idle entries, while updating each time its Bayesian beliefs on the future vector realizations. We formulate and study the CCS as a Restless Multi-Armed Bandit problem, generalizing the popular Cognitive Spectrum Sensing model, in which the CR can sense K out of the N sub-channels and propose a novel adaptive Finite Rate of Innovation (FRI) sampling method based on the CCS approach. While in general the optimum policy remains elusive, we provide sufficient conditions in which in the limit for large K and N the greedy policy is optimum. Numerical results corroborate our theoretical findings.

Original languageEnglish (US)
Article number7004089
Pages (from-to)1183-1198
Number of pages16
JournalIEEE Transactions on Signal Processing
Volume63
Issue number5
DOIs
StatePublished - Mar 1 2015
Externally publishedYes

Fingerprint

Innovation
Sampling

Keywords

  • Cognitive radio
  • compressive sensing
  • multi-channel sensing
  • myopic policy
  • opportunistic spectrum access

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Signal Processing

Cite this

The restless multi-armed bandit formulation of the cognitive compressive sensing problem. / Bagheri, Saeed; Scaglione, Anna.

In: IEEE Transactions on Signal Processing, Vol. 63, No. 5, 7004089, 01.03.2015, p. 1183-1198.

Research output: Contribution to journalArticle

@article{4b70ff755c1844e496a09452fa8b6776,
title = "The restless multi-armed bandit formulation of the cognitive compressive sensing problem",
abstract = "In this paper we introduce the Cognitive Compressive Sensing (CCS) problem, modeling a Cognitive Receiver (CR) that optimizes the K projections of a N > K dimensional vector dynamically, by optimizing the objective of correctly detecting the maximum number of idle entries, while updating each time its Bayesian beliefs on the future vector realizations. We formulate and study the CCS as a Restless Multi-Armed Bandit problem, generalizing the popular Cognitive Spectrum Sensing model, in which the CR can sense K out of the N sub-channels and propose a novel adaptive Finite Rate of Innovation (FRI) sampling method based on the CCS approach. While in general the optimum policy remains elusive, we provide sufficient conditions in which in the limit for large K and N the greedy policy is optimum. Numerical results corroborate our theoretical findings.",
keywords = "Cognitive radio, compressive sensing, multi-channel sensing, myopic policy, opportunistic spectrum access",
author = "Saeed Bagheri and Anna Scaglione",
year = "2015",
month = "3",
day = "1",
doi = "10.1109/TSP.2015.2389620",
language = "English (US)",
volume = "63",
pages = "1183--1198",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - The restless multi-armed bandit formulation of the cognitive compressive sensing problem

AU - Bagheri, Saeed

AU - Scaglione, Anna

PY - 2015/3/1

Y1 - 2015/3/1

N2 - In this paper we introduce the Cognitive Compressive Sensing (CCS) problem, modeling a Cognitive Receiver (CR) that optimizes the K projections of a N > K dimensional vector dynamically, by optimizing the objective of correctly detecting the maximum number of idle entries, while updating each time its Bayesian beliefs on the future vector realizations. We formulate and study the CCS as a Restless Multi-Armed Bandit problem, generalizing the popular Cognitive Spectrum Sensing model, in which the CR can sense K out of the N sub-channels and propose a novel adaptive Finite Rate of Innovation (FRI) sampling method based on the CCS approach. While in general the optimum policy remains elusive, we provide sufficient conditions in which in the limit for large K and N the greedy policy is optimum. Numerical results corroborate our theoretical findings.

AB - In this paper we introduce the Cognitive Compressive Sensing (CCS) problem, modeling a Cognitive Receiver (CR) that optimizes the K projections of a N > K dimensional vector dynamically, by optimizing the objective of correctly detecting the maximum number of idle entries, while updating each time its Bayesian beliefs on the future vector realizations. We formulate and study the CCS as a Restless Multi-Armed Bandit problem, generalizing the popular Cognitive Spectrum Sensing model, in which the CR can sense K out of the N sub-channels and propose a novel adaptive Finite Rate of Innovation (FRI) sampling method based on the CCS approach. While in general the optimum policy remains elusive, we provide sufficient conditions in which in the limit for large K and N the greedy policy is optimum. Numerical results corroborate our theoretical findings.

KW - Cognitive radio

KW - compressive sensing

KW - multi-channel sensing

KW - myopic policy

KW - opportunistic spectrum access

UR - http://www.scopus.com/inward/record.url?scp=84922788698&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84922788698&partnerID=8YFLogxK

U2 - 10.1109/TSP.2015.2389620

DO - 10.1109/TSP.2015.2389620

M3 - Article

AN - SCOPUS:84922788698

VL - 63

SP - 1183

EP - 1198

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 5

M1 - 7004089

ER -