Selective data collection in vehicular networks for traffic control applications

TL;DR

This paper presents a selective data collection method for vehicular sensor networks that reduces data transmission by transmitting only when traffic control decision uncertainty exceeds a threshold, improving traffic management efficiency.

Contribution

It introduces a novel uncertainty-based selective data transmission approach for VSNs, enhancing data efficiency in traffic control applications.

Findings

Significant reduction in data transmitted without compromising decision accuracy

Effective traffic flow prediction using online simulation techniques

Improved traffic control decisions at signalized intersections

Abstract

Vehicular sensor network (VSN) is an emerging technology, which combines wireless communication offered by vehicular ad hoc networks (VANET) with sensing devices installed in vehicles. VSN creates a huge opportunity to extend the road-side sensor infrastructure of existing traffic control systems. The efficient use of the wireless communication medium is one of the basic issues in VSN applications development. This paper introduces a novel method of selective data collection for traffic control applications, which provides a significant reduction in data amounts transmitted through VSN. The underlying idea is to detect the necessity of data transfers on the basis of uncertainty determination of the traffic control decisions. According to the proposed approach, sensor data are transmitted from vehicles to the control node only at selected time moments. Data collected in VSN are processed using on-line traffic simulation technique, which enables traffic flow prediction, performance evaluation of control strategies and uncertainty estimation. If precision of the resulting information is insufficient, the optimal control strategy cannot be derived without ambiguity. As a result the control decision becomes uncertain and it is a signal informing that new traffic data from VSN are necessary to provide more precise prediction and to reduce the uncertainty of decision. The proposed method can be applied in traffic control systems of different types e.g. traffic signals, variable speed limits, and dynamic route guidance. The effectiveness of this method is illustrated in an experimental study on traffic control at signalised intersection.