Localização no tempo e no espaço em redes de sensores sem fio

Abstract

Wireless Sensor Network (WSN) based applications are usually event driven. An event by itself can be defined as being composed of a causal criterion and a spatiotemporal criterion. The first criterion specifies the type of event, while the second specifies the location in time and space where the event occurred. A sensor node is able to identify the first criterion easily by using its own sensing devices. The spatiotemporal criterion, on the other hand, can only be identified when the sensor nodes of a WSN have synchronized clocks and are able to determine their physical location. Time and space information is also required by a number of algorithms and protocols in WSNs such as information fusion, object tracking, energy maps, and density control. Thus, synchronization and positioning for WSN-based applications are challenging problems that need to be addressed.In this thesis, we identify that synchronization and positioning in WSNs are actually two parts of the same problem, i.e., locating the nodes in a network in time-space. Neither location without time nor time without location is complete information in these networks. The similarities between the localization and synchronization problems also suggest that they can and should be addressed as a single problem. From our perspective, time can be seen as another dimension of space. As a consequence, we need to solve a 4D positioning problem. This approach allows us to save network resources and solve both problems more efficiently.We propose three different solutions for the localization in time and space problem that are suitable for different scenarios in WSNs. These proposed solutions are called the Synapse, Lightness, and Mobilis algorithms. Our proposed algorithms not only take advantage of the additional hardware resources required by the positioning mechanism in order to improve the performance and scalability of synchronization, but also benefit from the additional communication required by the synchronization mechanism to decrease positioning errors. We also present an extensive set of experiments to evaluate the performance of our algorithms. Our results indicate that our proposed schemes are suitable for implementation in WSNs and also highlight the advantages of solving both localization and synchronization problems with a unified algorithm.