An intelligent path recommendation protocol (ICOD) for VANETs

This work introduces a novel, intelligent, real-time, distributed, and flexible path recommendation protocol. We refer to the introduced protocol as an Intelligent path recommendation protocol (ICOD). This protocol aims to find the best path towards each destination in any grid-based layout area. The path is reactively reconfigured at each road intersection; this is based on the real-time traffic status of the Surrounding Road Segments (SRSs) of that intersection: i.e., any road segment that starts or ends with the road intersection. In order to handle the centralized behavior problems (i.e., bottleneck, single point of failure, etc.) which appeared in the previous path recommendation protocols, ICOD has constructed the path towards each destination in a distributed manner. Three different variants of ICOD are thus introduced to enable flexible selection of the best path towards each destination according to the following driver concerns and priorities: congestion avoidance, economical and context-aware. The congestion avoidance variant aims to reduce the traffic congestion, by recommending each vehicle to take the least congested path towards its destination. The economical variant recommends the best economical path in terms of fuel consumption and gas emission parameters of each candidate path. Finally, the context-aware variant considers the context of each traveled road segment while selecting the desired path. Thus, the located services and the road conditions are considered. We report on the performance of these different variants of ICOD and compare them with previous path recommendation protocols, using an extensive set of scenarios and experiments implemented by NS-2. From these experimental results, ICOD is shown to exhibit a good performance in terms of its ability to recommend the best path which decreases the delay of each vehicle by 50% in the process of obtaining the directions towards its destination; it also eliminates the bottleneck and single point of failure problems.