Safety- and QoS-aware management of heterogeneous sensor networks

Recently, we have proposed ANSWER: AutoNomouS Wireless sEnsor netwoRk as a service platform whose mission is to provide dependable information services to in-situ mobile users while satisfying their quality-of-service (QoS) requirements. Alongside with the stationary tiny sensors, the network employs more powerful mobile devices referred to as aggregation and forwarding nodes (AFNs). ANSWER exploits AFN mobility to support QoS requirements. However, as an AFN moves closer to an event (e.g., a hazardous spill), it may be at risk (e.g., due to potential damage). In this paper, we present a quantitative analysis of the interplay and balance between QoS support and asset safety. We propose a new scheme, called SAFER (for SAFEty-aware Relocation), which pursues relocation of the AFN in order to boost network performance without unnecessarily compromising AFN safety. SAFER uses historical data on detected events and employs an evolutionary neural network to assess the risk involved and predict good quality new position(s).