Automotive navigation systems are widely used by drivers to help them reach a particular destination. Mostly, such systems are designed to suggest routes which will help the driver to reach destination by covering less distance or consuming less time. This paper presents a technique that will select least congested path by using a vehicular traffic prediction technique that utilizes Global positioning data paired with Speed and Accelerometer Telemetry provided by a GPS unit, mounted on a vehicle, to improve the results of existing graph search methodologies currently being implemented to provide navigation data to users. Clustering of vehicles is done, based on similar positional and directional behavior. These clusters will have similar congestion levels. Traffic data identified through clustering is used to manipulate the path cost of the corresponding road, on an existing road network graph. It tries to improve real-time route suggestions by selecting less congested routes dynamically.

Clustering Dynamic Updates Path Cost Traffic Factor Route Select Algorithm