A practical data driven approach for the deployment of WSNs on realistic terrains

Abstract

Existing works on relay nodes placement in two-tiered wireless sensor networks (WSNs) assume two-dimensional terrains and ignore completely the impact of the topography on the communication quality. In contrast, this paper addresses constrained versions of the relay nodes placement problem while adopting more realistic assumptions. To ensure the connectivity of a two-tiered WSN deployed on a three-dimensional terrain, the proposed approach determines the relay nodes positions by analyzing the data extracted from the terrain topography. The main idea is to consider a concrete positions constraint, by requiring the placement of relay nodes only at the terrain crest points, which have a relatively wide visibility on the terrain and are well exposed to the sun. This allows minimizing the number of relay nodes required to ensure the connectivity, reduces the impact of the topography on the wireless communication quality, and offers the opportunity to equip the relay nodes with solar panels to extend their lifetimes. Based on this idea and inspired by the One-Step constrained Relay nodes Placement (OSRP) and Realistic constrained Relay nodes PLacement (RRPL) approaches, three algorithms, called 3D-OSRP (3D One-Step constrained Relay nodes Placement), 3D-RRPL-v1 (3D Realistic constrained Relay nodes PLacement version-1), and 3D-RRPL-v2 (3D Realistic constrained Relay nodes PLacement version-2), are designed, implemented, and evaluated by extensive simulations on realistic three-dimensional terrains. Obtained results demonstrate the relevance of considering positions constraint.Trans Emerging Tel Tech. 2019;30:e3558.wileyonlinelibrary.com/journal/ett network lifetime. Moreover, the limited communication range of the SNs requires the use of a multihop routing, in which the data passes through many intermediate nodes before reaching one of the CNs, which increases both packet loss probability and packet delivery delay. To overcome these challenges, recent works have suggested reinforcing this basic architecture, composed of SNs and CNs, by a third type of nodes, called relay nodes (RNs). [2][3][4] Having an extended energy autonomy 5,6 and a communication range often greater than that of SNs, 7,8 these RNs are designed to preserve or to restore the network connectivity by extensively participating in routing and driving data to the CNs, without contributing to the detection tasks assigned exclusively to the SNs. Nevertheless, the introduction of RNs generates new constraints that need to be taken into account. Indeed, the arrangement of these RNs in the RoI must be well examined according to several factors, which include, among others, the positions already occupied by the SNs, because the main role of the RNs is to reinforce or to restore network connectivity. In addition, because of their high-cost, 9 the number of RNs to be deployed must be minimized. Therefore, the deployment of RNs must be performed in a deterministic manner, 10 by precomputing their positions in the RoI, since the random deployment cannot respond ef...

Fault‐tolerant data transport backbone for 3D wireless sensor networks

“…Table 1 summarizes the proposed approaches for RN-based clustering of WSNs, for which we mention again the following shortcomings. First, except for some works that consider 3D RoI 13 or very specific RoI, 15,16 most existing approaches are designed and evaluated under the assumption that the RoI is 2D. Consequently, these approaches are inapplicable for a realistic 3D RoI.…”

“…Nevertheless, its extension for 3D RoIs is really difficult because the identification of the RNs positions is based on the Jarvis March algorithm, 26 which is a geometric method that is only applicable for specific or purely 2D surfaces. Similarly, our previous work 13 is the only one that takes into account the topography impact, by representing the 3D RoI by a Triangulated Irregular Network (TIN) model and considering that two nodes can communicate only if they are intervisible. However, as mentioned in Table 1, this work ensures only simple connectivity and ignores completely the fault-tolerance problem.…”

“…[9][10][11] Finally, the coverage and the connectivity become independent, where it is possible to satisfy them in two separate phases. 12,13 Nevertheless, although the RNs have relatively high energy resources, their permanent activities as NoA or NoT are a real concern, given the enormous data flow, they transport from the SNs to the CNs. The failure of some RNs can isolate a large number of SNs from the rest of the network.…”

“…In a recent work, 13 we have proposed a new RN-based clustering approach for 3D WSN, by taking into account the RoI topography impact. However, this approach targets only simple connectivity and does not deal with the aforementioned fault-tolerance problem.…”

Using Relay Nodes in Wireless Sensor Networks: A Review

“…In addition, these approaches suffer from a problem of practicalness, due to the omission of certain constraints having a quasi-permanent presence in real-life applications, or because of the adoption of unrealistic assumptions. For instance, excepting some works [3], [39], [30] that considered 3D or very specific RoI, all mentioned approaches are designed and/or evaluated under the assumption that the RoI is purely 2D, which makes them irrelevant for realistic 3D RoI, and limits their usage to very limited cases.…”