Energy-based localization in wireless sensor networks using second-order cone programming relaxation

Research output: Contribution to journalArticle

14 Citations (Scopus)


Source localization in wireless sensor networks (WSNs) aims to determine the position of a source in a network, given inaccurate position-bearing measurements. This paper addresses the problem of locating a single source from noisy acoustic energy measurements in WSNs. Under the assumption of Gaussian energy measurement errors, the maximum likelihood (ML) estimator requires the minimization of a nonlinear and nonconvex cost function which may have multiple local optima, thus making the search for the globally optimal solution hard. In this work, an approximate solution to the ML location estimation problem is presented by relaxing the minimization problem to a convex optimization problem, namely second-order cone programming. Simulation results demonstrate the superior performance of the convex relaxation approach. More precisely, the new approach shows an improvement of 20 % in terms of localization accuracy when compared to the existing approaches at moderate to high noise levels. Simulation results further show comparable performance of the new approach and the state-of-art approaches at low noise levels.

Original languageEnglish
Pages (from-to)1847-1857
Number of pages11
JournalWireless Personal Communications
Issue number3
Publication statusPublished - Aug 2014


  • Centralized localization
  • Energy-based localization
  • Maximum likelihood estimation
  • Nonconvex optimization
  • Noncooperative localization
  • Second-order cone programming (SOCP)
  • Semidefinite programming (SDP)
  • Wireless sensor networks

Fingerprint Dive into the research topics of 'Energy-based localization in wireless sensor networks using second-order cone programming relaxation'. Together they form a unique fingerprint.

  • Cite this