Propagation Model Alternatives for Outdoor Wireless Sensor Networks Sinan Kurt ASELSAN Inc., Ankara, Turkey Email: sinank@aselsan.com.tr Bulent Tavli TOBB-ETU, Ankara, Turkey Email: btavli@etu.edu.tr Abstract—Modeling propagation in Wireless Sensor Networks (WSNs) is a vital issue in system design and analysis. Energy efficiency, routing performance, Quality-of-Service, coverage and almost all aspects of WSNs are inherently related with the employed propagation model. Yet, in WSN research, even after more than a decade long continuing and widening research history, very simplistic and impractical propagation models are still utilized. In this paper, we briefly provide the necessary background on propagation modeling and concisely overview experimentally verified propagation models for WSNs. I. I NTRODUCTION Wireless sensor networks (WSN) are composed of a base station and a plurality of sensor nodes spatially distributed over a target region to remotely monitor desired phenom- ena. Most of the research activities on WSNs (e.g., routing, medium access control, position estimation, synchronization, lifetime optimization, modulation, error correction, security) are implicitly related to the propagation media and accurate propagation model of signals among sensor nodes. Accuracy and precision of the employed propagation model affects coverage area, selected transmission power, battery life of sensors, and consequently the lifetime of the network. In traditional wireless communications there exist well es- tablished and known propagation models that suit well to specific applications which are experimentally corrected and adopted to be used with high accuracy. Nevertheless for the WSN case, some of the models are used without awareness of the assumptions made in that model and consequently an unsuitable model could be used [1]. The aim of this paper is first to emphasize WSN specific constraints for a propagation model. Also available outdoor environmental WSN specific large-scale propagation model alternatives that are reported to be more accurate (justified by physical measurements) than widely used models are presented. Furthermore, application specific perspectives of these alternatives are discussed. II. WSN CONSTRAINTS WSNs have their own set of constraints which are not compatible with the assumptions made in most of the tra- ditional wireless communication systems. In other words, WSN specific propagation models require consideration of the constraints imposed by the nature of WSN use scenarios, as itemized below: 1) Low Antenna Heights: In most of the WSN deployment scenarios, a large number of small form factor sensor nodes are assumed to be placed on the ground. Obvi- ously, the antennas used in these devices are standing not more than a few centimeters above the ground. Although it is possible that the base station and/or some of the sen- sor nodes can be positioned at a higher elevation, most of the sensors are directly on the ground. Therefore, propagation models that assume antenna heights (both for the transmitter and the receiver) higher than a few meters are not directly applicable to WSNs. 2) Directivity of Antennas: The use of omnidirectional antennas are the usual assumption is WSNs. However, the assumption of a perfectly omnidirectional radiation pattern is not realistic for WSNs which are consisting of inexpensive tiny sensor platforms [2]. 3) Low Transmission Power: Transmission power level in WSN platforms are comparatively lower than many other wireless communication systems (i.e., in the order of 100 mW is a typical maximum transmission level in WSNs). Therefore, transmission ranges, typically, are upper limited by roughly 600 m for outdoor open fields which is much lower than the transmission ranges of most of the wireless communications systems (e.g., cellular systems). 4) Stationary network topology: Although mobile sensor nodes or mobile base stations are used in some WSN scenarios, in most WSN scenarios, sensor nodes are as- sumed to be stationary (not moving), therefore, channel models that capture the effects of relative mobility be- tween the transmitter and the receiver are not necessary in static WSNs (e.g., the impact of doppler effect is negligible). Furthermore, the level of mobility in WSNs with mobile nodes is not high (i.e., pedestrian mobility). III. WSN PROPAGATION MODEL ALTERNATIVES The most widely used propagation model for WSNs is free-space propagation model which can be simplified as: P r (d) = P0 d 2 , where P 0 is a reference power measured at a reference distance from transmitter. Simplicity of the model brings popularity. This model is used to account for energy dissipation of routing protocol simulations focusing on network lifetime and energy efficiency [3].