A hybrid intelligent Body Sensor Networks model
Thales Baliero Takao, Danilo Silva Ramos, J. O. Ferreira, Olegário Corrêa da Silva Neto, Adson Ferreira da Rocha, Talles M. G. de A. Barbosa · 2011
Body Sensor Networks (BSN) have gained interest in recent years from researches. Several promising prototypes are enabling many healthcare services. Despite of technological developments in sensing and monitoring devices, some issues related to BSN have still to be investigated. For example, information routing requires establishment of multi-hop paths which can be done considering the amount of consumed energy at each instant. Previous works addressed this issue by means of blind-routing algorithms such as Dijkstra. In this paper a Genetic Algorithm (GA) is evaluated in order to improve the system scalability. Another important aspect is related to autonomous processing capability expected of each sensor node. This work proposes and evaluates Self Organizing Map (SOM) as an unsupervised learning solution. Thus, the main purpose of this paper is to present a hybrid model combining GA and SOM in order to support intelligent BSNs. At the end of this paper a case study was evaluated in order to support a thermal biofeedback application.