No abstract

Wireless sensor networks can be used in habitat monitoring for detecting fire, in disaster for helping rescue teams and in agriculture for sensing humidity. Node localization is essential for some important sensor network applications. Despite the high relative accuracy of some localization algorithms, node localization is still an opened research area, due to the physical phenomena such as attenuation, reflection, diffraction, scattering and so forth. The current developed algorithms have different accuracy when are tested under dissimilar environments. We propose to use Smart Beacon Nodes (SBNs) to infer the Obstruction Level Indicator over an occupied area, then, use this indicator for estimating the distance among nodes. In our experimental simulation, SBNs decrease the node localization error of Triangular Centroid Localization and Weighted Centroid Localization up to 18%.

As far as we know, there is not a Node Localization Algorithm (NLA) that presents the same accuracy for all possible scenarios. We believe that a NLA should be able to "interpret" the dynamic information of the environment. In this sense, simple NLAs are rather focused and might perform well for specific scenarios and applications. Therefore, information fusion and context awareness seems to be an appropriate approach to address this issue. We propose the Smart Environmental Architecture for Node Localization (SEA-NL), which is composed by two main elements: (i) the Smart Beacon Nodes (SBNs) and (ii) the Logical Position of Nodes (LPN). In (i) the obstruction level indicator is estimated and can improve the estimation of distances among nodes. In (ii) environment information and a one to one relation between a node and an object are used andcan also improve location estimation. Via simulation, our architecture was tested indoors and outdoors considering three localization algorithms: the Weighed Centroid Localization (WCL), the Centroid Localization, and the Triangular Centroid Localization. Finally, we present an accuracy comparison among NLAs used in isolated way, and by using the SBNs, the LPN, and the SEA-NL, where our architecture improves WCL up to ~30.88%.