Ant-based vehicle congestion avoidance system using vehicular networks Mohammad Reza Jabbarpour a,n , Ali Jalooli a , Erfan Shaghaghi a , Rafidah Md Noor a , Leon Rothkrantz b , Rashid Hafeez Khokhar c , Nor Badrul Anuar a a Faculty of Computer Science & Information Technology, University of Malaya, Kuala Lumpur 50603, Malaysia b Faculty of Intelligent Systems, Delft University of Technology, Mekelweg 4, 2628CD Delft, The Netherlands c School of Computing & Mathematics, Charles Sturt University, WaggaWagga, NSW 2678, Australia article info Article history: Received 7 February 2014 Received in revised form 11 July 2014 Accepted 3 August 2014 Keywords: Vehicle traffic routing Ant colony optimization Vehicular networks Vehicle congestion problem Car navigation system abstract Vehicle traffic congestion leads to air pollution, driver frustration, and costs billions of dollars annually in fuel consumption. Finding a proper solution to vehicle congestion is a considerable challenge due to the dynamic and unpredictable nature of the network topology of vehicular environments, especially in urban areas. Instead of using static algorithms, e.g. Dijkstra and A*, we present a bio-inspired algorithm, food search behavior of ants, which is a promising way of solving traffic congestion in vehicular networks. We have called this the ant-based vehicle congestion avoidance system (AVCAS). AVCAS combines the average travel speed prediction of traffic on roads with map segmentation to reduce congestion as much as possible by finding the least congested shortest paths in order to avoid congestion instead of recovering from it. AVCAS collects real-time traffic data from vehicles and road side units to predict the average travel speed of roads traffic. It utilizes this information to perform an ant-based algorithm on a segmented map resulting in avoidance of congestion. Simulation results conducted on various vehicle densities show that the proposed system outperforms the existing systems in terms of average travel time, which decreased byan average of 11.5%, and average travel speed which increased by an average of 13%. In addition, AVCAS handles accident conditions in a more efficient way and decreases congestion by using alternative paths. & 2014 Elsevier Ltd. All rights reserved. 1. Introduction Over the last decade, vehicle population has dramatically increased all over the world. This large number of vehicles leads to heavy traffic congestion, air pollution, high fuel consumption and consequent economic issues (Narzt et al., 2010). In 2010, the American people faced a lot of difficulties due to vehicle congestion which forced their government to spend 101 billion dollars on the purchase of extra fuel (Schrank et al., 2012). Based on a report by Texas A&M Transportation Institute (Schrank et al., 2012), it is estimated that fuel consumption will rise up to 2.5 billion gallons (from 1.9 billion gallons in 2010) with a cost of 131 billion dollars in 2015. Accordingly, finding effective solu- tions with reasonable cost for congestion mitigation is one of the major concerns of researchers and industries in recent years. Building new, high-capacity streets and highways can mitigate some of the aforementioned problems. Nevertheless, this solution is very costly, time consuming and in most cases, impossible because of space limitations. On the other hand, optimal usage of the existing roads and streets capacity can lessen the congestion problem in large cities at a lower cost. Intelligent Transportation System (ITS) (Dimitrakopoulos and Demestichas, 2010) is a newly emerged system which collects real- time data for congestion monitoring using road side units (e.g. video cameras, radio-frequency identification (RFID) readers and induction loops) and vehicles as mobile sensors (i.e. in-vehicle technologies or smart phones). These data are used by car navigation systems (CNSs) to find the shortest path or optimal path from a source to a destination. Previous researches (Noto and Sato, 2000; Yue and Shao, 2007; Nazari et al., 2008) concentrated on using static algo- rithms (e.g. Dijkstra, 1959) and A n (Hart et al., 1968) to find the shortest path in CNSs. Conversely, current studies primarily focus on finding the optimal paths, considering various criteria by utilizing dynamic and meta-heuristic algorithms (Liu et al., 2007; Salehinejad and Talebi, 2008; Boryczka and Bura, 2013). This trend happens due to the dynamic nature of vehicular environments which depends on both predictable and unpredictable events and also because of the Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/engappai Engineering Applications of Artificial Intelligence http://dx.doi.org/10.1016/j.engappai.2014.08.001 0952-1976/& 2014 Elsevier Ltd. All rights reserved. n Corresponding author. Tel.: þ60 176238395. E-mail addresses: reza.jabbarpour@siswa.um.edu.my (M.R. Jabbarpour), ashkansp2@gmail.com (A. Jalooli), erfan_shaghaghi@siswa.um.edu.my (E. Shaghaghi), fidah@um.edu.my (R.M. Noor), l.j.m.rothkrantz@tudelft.nl (L. Rothkrantz), rkhokhar@csu.edu.au (R.H. Khokhar), badrul@um.edu.my (N.B. Anuar). Engineering Applications of Artificial Intelligence 36 (2014) 303–319