Ordenação de eventos e interpretação de contextos em redes de atuadores e sensores sem fio para a classe de aplicações de preparação para emergência
Visualizar/ Abrir
Data
2006-06-09Autor
Silva, Fernando Henrique dos Santos e
Metadata
Mostrar registro completoResumo
Wireless Sensor and Actor Networks (WSANs) are increasingly being deployed for fine-grain
monitoring of physical environments subjected to emergency situations such as fire, leaking
of toxic gases and explosions. In these environments a reliable interpretation of the collected
events is very important. However, the order in which events are received can interfere
directly with the correct interpretation of what is going on in the physical environment being
monitored. Therefore, the ordering of events in WSAN can be a major requirement for critical
security monitoring applications because by handling event ordering, ambiguities can be dealt
with, increasing the accuracy of what is being monitored. This work describes a novel
algorithm for event ordering in WSAN called OBC (Ordering by Confirmation). This low
latency and energy aware protocol was discussed and evaluated against existing solutions,
showing that it can save more energy and achieve lower latencies in situations where the
network density, event rate and network dimension increase. This work also presents a
context interpretation solution that can interpret from simple to complex contexts. In this
solution, actors are used to aggregate sensor events, eliminate ambiguities and redundancy
and realize context interpretations. Each actor of the WSAN is configured with rules that are
determined by the application in the network configuration phase. After ordered, correlated
events are interpreted resulting in contexts that go from simple questions, such as What is the
temperature of sector A? to more complex ones Why did the aircraft wing crack? . Context
information is exchanged among actors and sink(s) of the WSAN through the
publish/subscribe technique based on topics. Our solution was compared to a centralized
solution, in which events are collected by the sink. Results from simulations have shown that
by having actors for decentralized context interpretation and a fast energy-aware event
ordering algorithm, our solution can be a good alternative for emergency preparedness
applications where lives and patrimony are at risk.