BDBComp
Parceria:
SBC
Uma Abordagem de Fusão de Dados em Redes de Sensores para Reduzir o Impacto de Erros de Localização em Algoritmos de Rastreamento

Éfren L. SouzaEduardo F. NakamuraHorácio A. B. F. de Oliveira

In wireless sensor networks (WSNs), target tracking algorithms usuallydepend on geographical information provided by a localization algorithm.However, the errors introduced by such algorithms may affect the performanceof tasks that depend on information about node position. Information fusiontechniques are natural choices to try reducing such errors. In this paper, usethe Kalaman filter to reduce distance estimation errors used by localization algorithms:Recursive Position Estimation (RPE) and Directed Position Estimation(DPE). Then, we evaluate how two classical tracking algorithms (Kalmanand Particle filters) are affected by localization errors, when we use variousdistance estimates. Results show that using multiple distance estimates in thelocalization algorithms improves the tracking, but this feature should be usedwith caution, mainly, due to the associated communication cost.

http://www.lbd.dcc.ufmg.br/colecoes/sbrc/2010/0021.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD