Optimized data flow management in event driven wireless sensor networks: A case study on landslide surveillance

Abstract

In this paper, we are presenting a data flow and fault management architecture, developed for an event driven wireless sensor network used in landslide surveillance. Our concept system collects landslide data using a dedicated wireless sensor network. The WSN was configured with tree architecture and contains one gateway, relay nodes, and multiple end nodes. The WSN components were distributed in an array configuration inside the surveyed area. Each WSN component acquired landslide data by means of measurement points mounted on sensing rods and placed in the earth at specific depths. Relay nodes use local data gathered from multiple end nodes to compute a local data trend and evaluate possible alerts. The gateway acquires data from the WSN, manages the data flow and analyzes the global evolution of data in order to issue alerts and warnings related to landslide activity. Also, it verifies the responses of all landslide sensors in order to ensure data consistency and fault management in the entire network. Our proposed data flow management scheme is centered on the possible events that could drive the change of network response time. These events are handled by assessing local and global data trends with respect to predefined markers of landslide data. Once a local or global event is confirmed it triggers the adjustment of system response. Dynamic landslide events require a fast system response in order to ensure precise surveillance. The concept integrates a fault tolerant architecture that uses diagnostic fault codes to ensure data consistency and a flexible data path for all end nodes.