Consistent sensor, relay, and link selection in wireless sensor networks

Rocío Arroyo-Valles, Andrea Simonetto*, Geert Leus

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

6 Citations (Scopus)
27 Downloads (Pure)

Abstract

In wireless sensor networks, where energy is scarce, it is inefficient to have all nodes active because they consume a non-negligible amount of battery. In this paper we consider the problem of jointly selecting sensors, relays and links in a wireless sensor network where the active sensors need to communicate their measurements to one or multiple access points. Information messages are routed stochastically in order to capture the inherent reliability of the broadcast links via multiple hops, where the nodes may be acting as sensors or as relays. We aim at finding optimal sparse solutions where both, the consistency between the selected subset of sensors, relays and links, and the graph connectivity in the selected subnetwork are guaranteed. Furthermore, active nodes should ensure a network performance in a parameter estimation scenario. Two problems are studied: sensor and link selection; and sensor, relay and link selection. To solve such problems, we present tractable optimization formulations and propose two algorithms that satisfy the previous network requirements. We also explore an extension scenario: only link selection. Simulation results show the performance of the algorithms and illustrate how they provide a sparse solution, which not only saves energy but also guarantees the network requirements.

Original languageEnglish
Pages (from-to)32-44
Number of pages13
JournalSignal Processing
Volume140
DOIs
Publication statusPublished - Nov 2017

Bibliographical note

Accepted Author Manuscript

Keywords

  • Convex relaxations
  • Link and relay selection
  • Resources optimization
  • Sensor selection
  • Sparsity
  • Wireless sensor networks

Fingerprint

Dive into the research topics of 'Consistent sensor, relay, and link selection in wireless sensor networks'. Together they form a unique fingerprint.

Cite this