Optimizing communication and computational costs based on the edge participation probability and the evolved variational message passing algorithm in wireless sensor networks

• Published in: Ad hoc networks Vol. 101; p. 102093
• Main Authors: Hedhly, Kaouther, Laaraiedh, Mohamed, Abdelkefi, Fatma, Siala, Mohamed
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 15.04.2020
• Subjects: Edge participation probability technique; Evolved variational message passing; Position estimate; Evolved variational message passing; Position estimate; Edge participation probability technique
• ISSN: 1570-8705; 1570-8713
• DOI: 10.1016/j.adhoc.2020.102093

Cooperative localization has become an important issue for several applications. One of the most efficient techniques for cooperative localization is the Evolved Variational Message Passing (E-VMP) algorithm. This algorithm offers accurate position estimates with less computational complexity compared to other cooperative localization algorithms. What is more, communication cost is a challenging issue facing any localization technique which increases in loopy networks. Consequently, we propose in this paper a re-evaluated E-VMP version which can meet this challenge. We propound a specific Edge Participation Probability technique, namely EPP, for each edge in the considered network. This technique facilitates the distributed implementation without graph transformation which is highly adequate in loopy networks. Experiments and realistic simulations show that the proposed technique offers considerable localization accuracy for very low processing complexity and cost.