| Abstract |
Vehicle tracking systemswhich receive the location data from tracking devices fitted in vehicles, visualize the location of vehicles on a web map and generate reports based on the location data. Tracking applications are commonly used for various purposes such as Fleet Management, Asset Tracking, Transit Tracking, Fuel monitoring etc.The GIS views of fleet management applications are limited to view the vehicles of a single owner or organization based on the logged in user. For smart cities, it is highly essential to monitor the location of public transport vehicles to provide smart transportation facilities to the public. For smart city based vehicle tracking application, the officials have to locate all the public transport vehicles, which include buses, taxies and auto-rickshaw in the city and therefore the number of vehicles are huge compared to existing applications. Locations of vehicles are to be plotted on a web map using marker objects. In the existing systems, the markers for huge number of vehicles in the viewing area provide a crowded view, which is of no use to the end user and in the case of web based applications,it may cause browser to crash if the markers are created beyond the memory limit. This paper proposes a high-performancecluster based method to store, process and retrieve dynamic geospatial data using Redis, a memory based cache, to track and plot the vehicle locations as clusters to provide an informative view. This approach helps to store GPS data from large number of vehicles at sub second latency and fast retrieval of vehicles locations with the given geospatial bounds without persisting the data permanently. The proposed method is experimented and proved to be a faster system and gives an efficient view when compared to the existing ones. |
