Early Warning System for Landslides Using Wireless Sensor Networks

Abstract

Landslides occur in many areas in Sri Lanka, and they cause considerable damage to natural habitat, environment, economy and other resources. Monitoring, predicting and controlling are the three major challenges associated with landslides due to the randomness of the event. Yet, developing an accurate prediction mechanism with an effective early warning system has become a need of the hour since the damages and the losses caused by the landslides are intolerable. Although there are expensive and advanced mechanisms deployed in foreign countries to predict the possibility of occurring landslides, such as satellites and radar systems with artificial intelligence capabilities, Sri Lanka finds it difficult to afford them due to the high cost and the advanced technologies used. When compared with the existing high-end systems, a simple wireless sensor network which is capable of identifying the underground movements and soil conditions is a cost effective, practical solution. But, dealing with a large number of variables manually with no proper understanding about their contribution for the occurrence of a landslide is difficult. Machine learning, which is a method used to create complex models and algorithms that lend themselves to predict is a fruitful solution for that issue. This research work is carried out to develop a cost-effective early warning system for land slides using WSNs incorporating machine learning.