Offline Navigation:GPS based assisting system in Sathuragiri forests using Machine Learning

Abstract

Location aware services are the need of the hour in places where communication facilities are limited. The lack of locationbased services in dense forested areas with poor network connectivity has resulted in severe losses both human life and monetary in many a places around the world. The case of floods in Sathuragiri forest in Virudhunagar district of Tamil Nadu in May 2015 is a classic case of such deficiencies. This paper presents an efficient and robust offline based navigation system suitable for the Sathuragiri forest which can be customized for varying geographical regions. The dataset for Sathuragiri forest was downloaded from Google Maps and applied K-means Clustering to identify the possible centroids. The calculations were carried out for two cases namely plane surfaces and sloppy terrains wherein the altitude was also taken into consideration. A standalone GPS module with other accessories was designed to indicate the current position. Using Haversine Formula, the distance to all the nearby centroids are calculated and the system displays the path with minimum weight for the traversal. Various methods of clustering have been performed and desired result was achieved from K-means clustering