Learning-based Routing in Mobile Wireless Sensor Networks

  • Author(s) / Creator(s)
  • Technical report TR12-01. Limited energy supply is a chief concern when dealing with wireless sensor networks (WSNs). Thus, among other issues, routing protocols for WSNs should be designed with the goal of being energy efficient in the first place. For static networks this is already a challenge, given that different domains and application requirements lead to different con- straints, and it becomes an even more complex problem when nodes in the WSNs are mobile. In this paper we address this very problem and propose centralized and decentralized routing techniques that are help prolong the nodes lifespan. The main idea is to explore the movement patterns of the nodes, through simple but effective learning, in order to use the most effective and less costly routing paths. Our experiments, using a real dataset, show that our proposed decentralized approach is twice as more energy-efficient than the centralized one, while being only marginally less effective. In addition, both outperform the well-known AODV protocol for ad-hoc mobile networks. | TRID-ID TR12-01

  • Date created
  • Subjects / Keywords
  • Type of Item
  • DOI
  • License
    Attribution 3.0 International