Threshold Based Data Aggregation Algorithm To Detect Rainfall Induced Landslides.
Maneesha Vinodini Ramesh, P. V. Ushakumari · International Conference on Wireless Networks · 2008
Abstract Landslides are one of the environmental disas-ters that cause massive destruction of human lifeand infrastructure. Real time monitoring of a land-slide prone areas are necessary to issue fore warn-ing. To accomplish real time monitoring, massiveamount of data have to be collected and analyzedwithin a short span of time. This work has de-veloped a method for effective data collection andaggregation by implementing threshold alert lev-els. The sampling rates of threshold alert levelswill determine amount of data collected and aggre-gated which will reduce the power consumption byeach wireless sensor nodes. This work also helps todetermine the appropriate sampling rates for eachthreshold level, and the expected number of datapackets in the queue. The time delay in receivingthe data packet at the analysis station can be de-termined by using the value of expected number ofdata packets in the queue. 1 Introduction Environmental disasters are largely unpre-dictable and occur within very short spans of time.Fore warning of environmental disasters is morechallenging than other applications due to the factthat they occur within a very short span. Thereforetechnology has to be developed to capture relevantsignals in a very short span of time. Wireless sen-sors are one of the cutting edge technologies thatcan respond to rapid changes of data.The fore warning of an environmental risk canbe done if and only if the relevant data within veryshort time spans are quickly captured with hightime resolution, processed and transmitted to theanalysis station. Wireless sensor network technol-ogy has all the above mentioned capabilities. How-ever, it has its own limitations such as relatively lowamounts of battery power and low memory avail-ability compared to many existing technologies. Itdoes, though, have the advantage of deploying sen-sors in hostile environments with a bare minimumof maintenance. This fulfills a very important needfor any real time monitoring.The scenario under consideration for this workis landslides. We aim to deploy wireless sensornetworks in a landslide prone area to detect land-slides