A Predictive and Intelligent Environmentally Friendly Vehicle Navigation System

Abstract

The field of Intelligent Transportation System (ITS) has been experiencing a remarkable growth in a wide range of vehicle applications research, categorized into safety to reduce or eliminate crashes, mobility to mitigate congestion and environmental impact, and convenience to provide driver connection to media and social network services. This research focuses on the mobility application and aims to provide drivers the least congested transportation route choices enabled by the ITS Vehicle to Vehicle (V2V), Vehicle to Infrastructure progressive vehicle navigation system.

Recent research in vehicle navigation systems has proposed energy consumption/emission optimized routing methodologies using historical traffic data modeling. More than 50% of congestion in U.S. cities is nonrecurring congestion [1]. Nonrecurring congestion reduces the availability of the traffic network, thus rendering historical traffic data-based systems insufficient in more than 50% of the cases. Real-time traffic data modeling provides an enhanced performance in traffic congestion assessment; however, greater performance is expected with a predictive traffic congestion model with increased certainty.

This research starts by reviewing the conventional shortest path and fastest path vehicle routing methodologies. It also reviews key search algorithms, namely the well-known unidirectional search algorithm Dijkstra and the bidirectional search algorithm A*.

Having presented the enhanced performance of eco-routing, the research introduces the predictive traffic information assessment and integration approach. Additionally, the approach aims to offer the driver the flexibility to optimize travel costs such as energy consumption, emission and travel time. The assessment of predictive traffic information modeling using wireless communication data has been limited due to the difficulty in objectively and quantitatively evaluating energy and emission reduction effects using ITS technology. The capabilities of Petri Net extend beyond other similar mathematical modeling languages, such as neural networks, to include analysis control and graphical representation. It is natural to model the optimal problem on a cost-dependent petri net graph where the travel costs are: energy consumption, emission and travel time. We propose an algorithm based on Dijkstra’s unidirectional search algorithm and introduce speedup techniques that may be applied to the cost-dependent network. Our methodology deals efficiently with the accuracy of the solution in a dynamic environment where selective travel cost is dynamically updated to enable optimal route solution.