DyReT: A Dynamic Rule Framing Engine Equipped With Trust Management for Vehicular Networks

Abstract

Managing a dynamic traffic system is a challenging task in vehicular environments. Clarity of vehicular data for efficient decision making is vital in Intelligent Transportation Systems (ITS). Huge volumes of vehicular data are collected and processed during vehicular transactions. Pre-processing the huge amounts of raw vehicular data followed by framing effective traffic rules to take appropriate rapid decisions by the ITS on the vehicles continues to be a challenging problem. Most of the current studies done on ITS have proposed decision making strategies to handle only specific vehicular events and many lacked framing intelligent dynamic decision rules along with appropriate actions, representing all traffic events prevailing in the vehicular environment. This study proposes a versatile decision engine implanted with a two-stage mechanism. In the first stage, we propose a novel data cleaning algorithm to identify and remove dirty data from the voluminous vehicular dataset. In the second stage, a unique rule framing mechanism is suggested to frame dynamic traffic rules along with their actions using real-time vehicular data. The vehicular entities take suitable decisions to respond to the traffic events based on these rules and their associated actions. A new Naïve Bayesian classifier is proposed in this study to test the new rule framed with the trained rules set, either to accept or reject the new rule for further processing. The algorithms are developed and implemented using machine learning concepts. Experimental and comparative analysis was carried out with other related referred studies to evaluate the performance of the proposed algorithms. Although the proposed decision engine is generic enough for decision making in most ITS use-cases, discussion in this article elaborates on its applicability in use-cases provisioning trust management.