Publication

Abstract : Landslide prone areas are mostly in remote regions, with very limited network connectivity. Hence it is very challenging to develop a continuous monitoring system, which can deliver early warning of landslides. For developing most appropriate communication architecture we need to consider the following factors that could deliver long term monitoring and real-time early warning of landslides. The factors are: (1) frequency of data collection from spatially distributed heterogeneous sensors based on their impact on the landslide initiation (2) acceptable tolerance limit of latency for each type of data packet arrival, (3) adaptive bandwidth requirement for efficient data transfer with respect to balance energy in each wireless sensor nodes, (4) adaptive routing of the data based on the propagation, terrain and climatic effects, (5) remote maintenance using node level reconfiguration and network level reconfiguration (6) secured real time data transfer (7) scalable to multi-site deployments etc. In this work, considering all the above factors we have developed context aware heterogeneous communication architecture. We have deployed the proposed communication architecture in two landslide prone areas, where one of the architecture is functional for the last ten years. This architecture has supported in collecting real time data from more than 150 geophysical sensors in adaptive frequency rate, remote configuring the sensor sampling rate, remotely triggering new software updates, providing prioritized service delivery based on the landslide alert level, data dissemination based on the warning levels etc. We have also designed and developed a Lightweight Management Framework (LMF) for this real-time, 24/7 operational, heterogeneous network. This LMF provides the ability to incorporate different heterogeneous networks such as 802.15.4, 802.11b/g/n, VSAT, GPRS, GSM, Internet and also proprietary wireless sensor network and hardware architectures. It also handles various network failures, data corruption, packet loss, and congestion problems. In this paper, we will discuss the performance evaluation and validation results of this architecture for achieving real-time monitoring and warning of landslides.