Dergi makalesi Açık Erişim

Optimal path selection under different data redundancy policies for energy efficient wireless sensor networks

2016    Altin-Kayhan, Aysegul; Sendil, Melek

A wireless sensor network (WSN) consists of energy-limited autonomous tiny devices spatially distributed so as to monitor their environments. They have a very wide application domain ranging from military applications to health care services, logistics, or agriculture. Efficient use of sensor energies and network security are of utmost importance, given their possible application domains and operational properties. Multipath routing and multi-copy forwarding are well-known approaches shown to be useful against several threats. In this paper we study the design of an energy-efficient and reliable WSN. We discuss four different routing strategies with multipath routing and various levels of multi-copying, namely, the no, full, and partial redundancy cases. Moreover, we incorporate balanced energy dissipation in our models not only to improve network lifetime, which is the time until the first sensor in the network exhausts its energy, but also to eliminate critical nodes, which are appealing for malicious attacks. We construct mathematical models of all these protocols and present some test results so as to display the trade-off between energy efficiency and network reliability. We compare their performance in terms of network lifetime and the amount of data successfully transmitted in case of an inconvenience. The full data redundancy protocol yields a more reliable design than the balanced utilization with no redundancy or the partial redundancy protocols in exchange for a quite significant degradation in network lifetime. However, the partial redundancy protocol provides a slightly less reliable WSN than the full redundancy case with almost 60% longer lifetime. As a result, we can suggest that our partial multi-copy protocol with extended balanced utilization rule is quite favourable for operating networks energy efficiently and securely.