IFAC-PapersOnLine 49-3 (2016) 249–254 ScienceDirect ScienceDirect Available online at www.sciencedirect.com 2405-8963 © 2016, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. Peer review under responsibility of International Federation of Automatic Control. 10.1016/j.ifacol.2016.07.042 © 2016, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. Keywords: Eco-Speed Control; Connected Vehicle; Fuel Consumption; Vehicle Dynamics Model; Signalized Intersection. 1. INTRODUCTION With the development of information and communication technology, connectivity between vehicles and between vehicles and transportation infrastructure was made possible. For instance, information of signal phasing and timing (SPaT), location and speed of vehicles could be easily transmitted and exploited for any application. Studies showed that vehicle fuel consumption levels in the vicinity of signalized intersections are dramatically increased due to vehicles’ deceleration and acceleration (Barth et al., 2008; H. Rakha et al., 2003). During the past decades, many studies have focused on changing traffic signal timings to optimize vehicles’ delay and fuel levels (Li et al., 2004; Stevanovic et al., 2009). Recently, researchers attempted to use connected vehicles and infrastructure technologies to develop eco- driving strategies that are more fuel-efficient. One of such applications is the Eco-Speed Control (ESC) which was developed to optimize individual vehicle fuel consumption by recommending a fuel-efficient trajectory using advanced information from surrounding vehicles and upcoming signalized intersections (Rakha et al., 2012). Various ESC algorithms were developed in recent years. Malakorn and Park proposed a cooperative adaptive cruise control system by using SPaT to minimize absolute acceleration levels of vehicles and reduce fuel consumption level (Malakorn et al., 2010). Kamalanathsharma and Rakha developed a dynamic programming based fuel-optimization strategy using recursive path-finding principles, and evaluated the developed strategy using an agent-based modelling approach (R. Kamalanathsharma et al., 2014). Asadi and Vahidi proposed a schedule optimization algorithm to allocate “green-windows” for vehicles to pass through a series of consecutive signalized intersections (Asadi et al., 2011). Most of ESC algorithms are developed and tested in a traffic simulation environment where vehicles are forced to follow the recommended speed as calculated by the ESC algorithms. However, many problems that are not treated in simulation software need to be solved in order to implement ESC in the field, such as communication latency, system malfunction, data collection error, driver perception/reaction delay, driver distraction resulting from following posted recommended speed, etc. Few studies attempted to investigate the potentials of implementing ESC in the field. For instance, Barth and Xia developed a dynamic eco-driving system and conducted a field test on arterial roads (Barth et al., 2011; Xia, 2014). However, vehicle fuel consumption is not explicitly considered in their algorithm objective function. Instead, their algorithm attempts to optimize vehicle acceleration and deceleration profiles to minimize the total tractive power demand and the idling time so that the fuel consumption levels are also reduced (Barth et al., 2011). This paper describes the preliminary field-controlled tests of an ESC algorithm developed to provide a “fuel-optimized” speed profile from upstream to downstream of a signalized intersection. In the tested algorithm, minimizing the fuel consumption level from upstream to downstream of the intersection is set as the objective function, and various constraints are constructed using the relationship between Abstract: This paper develops and addresses the implementation issues associated with the field application of an Eco-Speed Control (ESC) system that computes and recommends a fuel-efficient trajectory to drivers using signal timing and phasing data received from downstream-signalized intersections. From an algorithmic standpoint, the proposed process addresses the possible scenarios that a driver may encounter while approaching a signalized intersection. Alternatively, from an implementation standpoint the paper overcomes the challenges associated with communication latency, data errors, real-time computation, and smoothness of the drive in developing the system. The Virginia Smart Road connected vehicle controlled facility at the Virginia Tech Transportation Institute (VTTI) was used to conduct a preliminary proof-of-concept testing of the proposed ESC system. The testing included driving on downhill and uphill approaches for four red indication offset values. In total 192 trips were conducted using four different participants. The analyzed data indicate that the proposed system is very promising, producing an average reduction in fuel consumption levels and travel times in the range of 17.4 and 8.4 percent, respectively. *Virginia Tech Transportation Institute, Blacksburg, VA 24061 USA (e-mail: hchen@vtti.vt.edu; hrakha@vtti.vt.edu; IEl-Shawarby@vtti.vt.edu; almannaa@vt.edu ). ** Université de Tunis El Manar, Ecole Nationale d'ingénieur de Tunis, Tunis, Tunisia (e-mail: amlouliz@vtti.vt.edu). Hao Chen*, Hesham A. Rakha*, Amara Loulizi**, Ihab El-Shawarby* and Mohammed H. Almannaa* Development and Preliminary Field Testing of an In-Vehicle Eco-Speed Control System in the Vicinity of Signalized Intersections