Uma rede de sensores sem fio multicamada, multimodal para monitoração ambiental

Abstract

The evolution ofWSN technology observed over the last years, allowed the development of new sensor node platforms employing the most different sensing modalities, specially visual sensing with video cameras. This new sort of devices culminated with WSNs that use heterogeneous elements organized into node tiers with different sensing units. These nets, called Multitier, Multimodal Wireless Sensor Networks (WSN M2), are the object of study of this work. Dissertation objectives include the design, modeling, implementation and behavior analysis of a WSN M2 for environmental monitoring. The proposed network employs three sensor node profiles hierarchically distributed into tiers that collect data from the environment using two distinct sensor modalities: infrared radiation and visual information. A visual tracking application is defined as a proof of concept for the WSN M2. During network design and modeling aspects such as sensor node's resources intelligent usage were considered. Sensing, processing and disseminating services were configured to work on demand, that is, the sensor nodes only executed these services whenever they were necessary to application. This behavior is the best fit for a visual sensing application where nodes sensing and processing energy expenses have non negligible costs. Experimental results obtained thought WSN simulation show that the multi-tier, multimodal approach is at least 2.2 times more economic concerning energy expenses when compared to a muti-tier network with a single sensing modality and at least 11 times more economic when compared to a single-tier homogeneous network. This energy savings however, do not sacrifice the WSN M2 sensing quality. A network prototype using commercial off-the-shelf electronic components, sensors and node platforms was also developed. This prototype demonstrates the viability of the WSN M2 technology for environmental monitoring.