Método de multilateração para algoritmos de localização em redes de sensores sem fio

• Autor(a) principal: Müller, Crístian
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Manancial - Repositório Digital da UFSM
• dARK ID: ark:/26339/0013000008f83
• Texto Completo: http://repositorio.ufsm.br/handle/1/5456
• topic: Redes de sensores sem fio; Algoritmos de localização; Multilateração; Máxima verossimilhança; Wireless sensor network; Location algorithms; Multilateration; Maximum likelihood; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

The wireless sensor networks have made great progress in the last decade and are increasingly present in several fields like security and monitoring of persons, animals or items, medicine, military area and many others that, due to technological developments, have become viable. These networks are formed by sensor nodes that have extremely limited resources, such as processing and data storage capacity, data transmission rate and energy available for operation. Thus, in networks with hundreds or even thousands of nodes, it is unfeasible to locate each one of them with global positioning devices, because those will considerably increase the cost and power consumption. As localization knowledge by the nodes is required in applications such as tracking, monitoring and environmental data collection, location algorithms were created to cheapen and/or improve this task. Thus, this master thesis presents the development of a low complexity iterative multilateration method, since most location algorithms uses some kind of multilateration. To prove this new method efficiency, a simple simulator based on the Matlab software was created in order to evaluate, in terms of location error, accuracy and robustness in a scenario with random arrangement of the reference nodes, log-normal propagation model with shadowing and received signal strength distance estimation. Under these conditions, the presented multilateration method presents inconsiderable loss of accuracy in comparison to the maximum likelihood method and also a low number of iterations is required. In this way was possible to increase the location algorithms accuracy without this entailing an increase in complexity and power consumption.