Fading-aware reliable and energy efficient routing in wireless sensor networks Janos Levendovszky a , Long Tran-Thanh a, * , Gergely Treplan b , Gabor Kiss a a Budapest University of Technology and Economics, Department of Telecommunications, H-1117 Magyar tud. krt. 2, Budapest, Hungary b Faculty of Information Technology, Peter Pazmany Catholic University, H-1083 Práter u. 50/A Budapest, Hungary article info Article history: Available online 15 July 2010 Keywords: Wireless sensor networks Reliability Fading Energy awareness Routing abstract In this paper, we introduce two fading-aware reliability based routing algorithms for wireless sensor net- works (WSNs) with lossy radio links. The proposed algorithms are able to find optimal multi-hop paths in polynomial complexity, over lossy links, which are modeled by using standard fading models (e.g. Ray- leigh and Rice fading). These algorithms minimize the energy consumption and ensure reliable packet transmission to the base station (BS) at the same time. A reliable path is defined in terms of successful packet transfer to the BS despite the lossy links. More precisely, the probability of correct reception of the packet at the BS must exceed a predefined threshold. The first algorithm minimizes the total energy consumption sending a packet over the selected path to the BS. On the other hand, the second algorithm selects a path which maximizes the minimum remaining energy on the node closest to exhaustion and, as a result, balances the energy consumption yielding high longevity. In both cases, reliable and energy effi- cient packet forwarding in WSN can be reduced to a constrained optimization problem. By using a spe- cific link metrics, these problems can then be mapped into shortest path problems solved in polynomial time. Thus the obtained results ensure the selection of reliable paths which also guarantee minimum energy consumption in real time. Ó 2010 Elsevier B.V. All rights reserved. 1. Introduction Due to the recent advances in electronics and wireless commu- nication, the development of low-cost, low-energy, multifunctional sensors have received increasing attention [1,2]. These sensors are compact in size and besides sensing they also have some limited signal processing and communication capabilities. However, these limitations in size and energy make the wireless sensor networks (WSNs) different from other wireless and ad hoc networks [3]. As a result, new data packet transmission methods must be developed with special focus on energy effectiveness in order to increase the lifetime of the network which is crucial in case applications, where recharging of the nodes is out of reach (e.g. military field observa- tions, living habitat monitoring, etc., for more details see [2]). Although a number of methods has been developed for energy aware data packet transmission in WSNs, such as destination- sequenced distance-vector (DSDV) routing [4], dynamic source routing (DSR) [5], and ad hoc on-demand distance-vector (AODV) routing [6], much of the research works is based on idealized assumptions about the wireless channel characteristics. That is, wireless communication can be perfect in term of packet loss within a circular radio range. However, several recent studies have convinced researchers that there is a need to replace this idealistic channel model with a more realistic one [7]. To date, a number of methods have been proposed in order to deal with realistic wireless communication models. In these meth- ods, the possible forwarding nodes are typically carefully evaluated and the node of a higher probability of delivery is then selected as a forwarding node. However, the applied evaluation metrics vary in different approaches. For instance, in GeRaF [8] the geographic dis- tance and a loss-aware metric in ETX [9] was used. Furthermore, Zamalloa and Krishnamachari proposed a position-based routing method using metrics similar to ETX [7]. These works, however, are typically failed in accurately modeling the reliability of wireless communication, since they usually use simplified channel models. Against this background, using fading models for wireless com- munication, such as Rayleigh and Rice, [10], this paper addresses reliable packet transmission in WSN when packets are to be received on the base station (BS) with a given reliability in terms of keeping the transmission error probability under a given thresh- old. Since in realistic communication channel models, the success of every individual packet transmission depends on the distance and the power of transmission, the probability of correct reception will diminish exponentially with respect to the number hops, in the case of multi-hop packet transfers. In this paper, we propose two new optimal algorithms which have minimal energy consumption subject to the constraint of guaranteeing reliable packet transfer to the BS. Within the first 0140-3664/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.comcom.2010.07.005 * Corresponding author. Tel.: +36 302074378. E-mail addresses: levendov@hit.bme.hu (J. Levendovszky), ttl@hit.bme.hu (L. Tran-Thanh), trege@digitus.itk.ppke.hu (G. Treplan), kissg@hit.bme.hu (G. Kiss). Computer Communications 33 (2010) S102–S109 Contents lists available at ScienceDirect Computer Communications journal homepage: www.elsevier.com/locate/comcom