IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 64, NO. 5, MAY 2015 1973 Speed Adaptive Probabilistic Flooding for Vehicular Ad Hoc Networks Yiannos Mylonas, Member, IEEE, Marios Lestas, Member, IEEE, Andreas Pitsillides, Senior Member, IEEE, Petros Ioannou, Fellow, IEEE, and Vicky Papadopoulou Abstract—A significant issue in vehicular ad hoc networks (VANETs) is the design of an effective broadcast scheme that can facilitate the fast and reliable dissemination of emergency warning messages in the vicinity of an unexpected event, such as a traffic accident. In this paper, we propose a novel solution to this problem, which we refer to as speed adaptive probabilistic flooding. The scheme employs probabilistic flooding to mitigate the effects of the broadcast storm problem, which is typical when using blind flooding, and its unique feature is that the rebroadcast probability is adaptively regulated based on the vehicle speed to account for varying traffic densities within the transportation network. The motivation behind this choice is the identification of the exis- tence of phase transition phenomena in probabilistic flooding in VANETs, which dictate a critical probability being affected by the varying vehicle traffic density and are shown to be linearly related to the vehicle speed (a locally measurable quantity). The scheme is evaluated using simulations on different sections of the freeway system in the City of Los Angeles, CA, USA. Simulation results indicate that the proposed scheme fulfills its design objectives, as it achieves high reachability and low latency of message delivery with low overhead in a number of representative scenarios. The scheme is also shown to outperform existing solutions, including Global-Positioning-System-based, and exhibits robustness with re- spect to different road topologies and parameters such as the transmission range of vehicles and the number of hops. Index Terms—Broadcasting, vehicular ad hoc networks, vehicle safety, wireless communication. I. I NTRODUCTION T HE rapid growth in transportation needs and car owner- ship has led to a severe increase in road traffic world- wide, which has generated a number of social and economic problems: congestion of road transportation networks, environ- mental hazards, and, above all, traffic accidents. In 2008 alone, about 6 million traffic accidents in the United States accounted for 37 231 fatalities, 2.35 million injuries, and $230 billion in Manuscript received July 12, 2012; revised January 11, 2013, September 20, 2013, and February 4, 2014; accepted June 15, 2014. Date of publication July 15, 2014; date of current version May 12, 2015. This work was supported by the Emergency Vehicular ad hoc networks (EM-VANETS) project. Y. Mylonas and A. Pitsillides are with the Department of Computer Science, University of Cyprus, 1678 Nicosia, Cyprus (e-mail: mylonasy@cs.ucy.ac.cy; andreas.pitsillides@ucy.ac.cy). M. Lestas is with the Department of Electrical Engineering, Frederick University, 1036 Nicosia, Cyprus (e-mail: eng.lm@frederick.ac.cy). P. Ioannou is with the Electrical Engineering Department, University of Southern California, Los Angeles, CA 90089 USA (e-mail: ioannou@usc.edu). V. Papadopoulou is with the Department of Computer Science and Engineering, European University Cyprus, 2404 Nicosia, Cyprus (e-mail: v.papadopoulou@euc.ac.cy). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TVT.2014.2339316 damaged property [1]. Government agencies and automotive industries are responding by investing billions of dollars in an effort to reduce these terrifying numbers. Intelligent trans- portation systems (ITS) [2] are considered central cornerstones in this effort, and vehicular ad hoc networks (VANETs) are emerging as the preferred network design for ITS technologies. The 802.11p standard [3], which is part of the IEEE WAVE protocol stack, supports both vehicle-to-vehicle and vehicle- to-infrastructure communications, allowing the formation of VANETs, which are envisioned to accommodate the new gen- eration of cooperative safety applications. VANETs can extend the information horizon of the drivers, and cooperative hazard warning applications may utilize this spatially broader view of the surrounding environment to alert drivers of potentially dangerous situations at an earlier stage. In case of an unex- pected event such as a traffic accident, a weather hazard, or a road works hazard, a vehicle that is appropriately equipped to detect the event will become an abnormal vehicle (AV) and utilize the underlying VANET to issue emergency warning messages (EWMs) to all neighboring vehicles warning them of the imminent danger. A major challenge in such a scenario is the design of the information dissemination scheme that will facilitate the reliable and low-latency transfer of the EWM to all vehicles in the vicinity of the unexpected event. It is critical that all vehicles within the area of interest receive the EWM with high probability, since a single uniformed vehicle can cause a traffic accident, and it is also significant that the transfer is completed with the minimum possible delay to give additional time to the driver or the automated collision avoidance system to respond to the potential danger and improve the safety level on the road. A straightforward solution to the aforementioned problem is blind flooding [4], which is a scheme that involves each vehicle rebroadcasting the EWM whenever it receives it for the first time. However, the ability of blind flooding to fulfil the design objectives is challenged by stressed communication channel conditions in cases of high vehicular traffic densities. Blind flooding is known to effectively work in sparse and moderately dense networks; however, its performance significantly de- grades in highly dense networks where a large number of redun- dant messages are generated. These redundant messages lead to unnecessary message collisions, increased contention, and high latency, which challenges the stringent delay requirements of the considered application. The problem is widely known as the broadcast storm problem [4], and a number of solutions have been proposed in the literature to mitigate its effects [5]– [14]. The main idea has been to reduce the number of nodes 0018-9545 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.