IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, VOL. 9, NO. 4, DECEMBER 2008 625 Capability-Enhanced Microscopic Simulation With Real-Time Traffic Signal Control Fang Clara Fang and Lily Elefteriadou Abstract—An application programming interface (API) is a feature that is available in some traffic simulation programs to enhance their capabilities by allowing users to customize changes in simulation such as driver behaviors, vehicle characteristics, user-defined control strategies, and advanced Intelligent Trans- portation Systems (ITS) applications. This paper presents an API in AIMSUN, which is a stochastic and microscopic simulation model, to evaluate a novel real-time signal control technique based on the dynamic programming (DP) algorithm. A transportation network of diamond interchanges is first created and calibrated in the AIMSUN environment. The API, which creates a dynamic link between the DP algorithm and AIMSUN, is then developed and deployed in C++. During simulation runtime, real-time traffic measurements, including vehicle counts and speeds, are provided by detectors in the network and fed into the DP algorithm that subsequently makes a decision on a signal control plan. The signal plan is then transferred back to and implemented in the simulated network, which emulates its actual operation. Extensive simulations have shown that the new signal control technique is superior to other common offline signal optimization tools in terms of handling the demand fluctuations. This paper has demonstrated that the API function is a useful tool to assess new ITS applications that are unavailable in simulation programs. Index Terms—Adaptive control, adaptive systems, application programming interface, intelligent systems, real time systems, simulation, traffic control (transportation), transportation. I. I NTRODUCTION I NTELLIGENT Transportation Systems (ITS) encom- passing a broad range of electronics technologies and communications-based information technology have been in- tegrated into our transportation system’s infrastructure and in vehicles to improve safety and relieve congestion. Computer simulation is a valuable tool to evaluate new ITS-oriented oper- ational strategies and algorithms before they are implemented in the field. However, it has been challenging as most of the in- novative ideas, such as adaptive ramp metering strategies, real- time adaptive signal control, Global Positioning System-based route guidance, and intelligent freeway incident management [1]–[4], are still not available in popular simulation models. Manuscript received August 28, 2007; revised March 18, 2008, May 29, 2008, and June 18, 2008. First published November 7, 2008; current version published December 1, 2008. The Associate Editor for this paper was M. Brackstone. F. C. Fang is with the Department of Civil, Environmental, and Biomedical Engineering, College of Engineering, Technology, and Architecture, University of Hartford, West Hartford, CT 06117 USA (e-mail: fang@hartford.edu). L. Elefteriadou is with the Transportation Research Center, Department of Civil and Coastal Engineering, University of Florida, Gainesville, FL 32611 USA (e-mail: elefter@ce.ufl.edu). 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/TITS.2008.2006770 Consequently, they could only be simulated by means of an application programming interface (API) module provided by simulators such as Paramics [5] and AIMSUN [6]. An API serves as a data exchange interface between user-defined ap- plications and the simulation environment. An API’s function is generally twofold: 1) to override the simulator default models, such as car-following and lane- changing models, and 2) to get access to external comple- mentary modules. A complementary module could be any ITS application, for example, advanced ramp metering or incident management [7]. So far, nearly all published research on the use of API is within Paramics, a microscopic traffic simulation tool, to test some traffic control strategies. The California Part- ners for Advanced Transit and Highways (PATH) project team [7]–[9] was perhaps the first group to use API functions in Paramics. They developed plug-in API modules to modify ele- ments of driver behaviors for modeling high-occupancy vehicle lane operations on I-680 and to test actuated signal control, time-based ramp meter control, and path-based routing. The research has been disseminated during ITS meetings [10], [11]. Followed by PATH’s work, a few other researchers have also applied API in Paramics for various purposes. Bartin et al. [12] implemented a gap acceptance and rejection binary probit model based on field data for the Collingwood and Brooklawn circles in New Jersey. Differences between simulation results of the API-enhanced model and the default Paramics model were presented. Park et al. [13] developed an API for coordinated actuated signal control logic for the Paramics microscopic simulation program. The control allows phase skip, gap out, and force off, which are generally not supported by the existing functions in Paramics. Smith et al. [14] have written an API module for simulating toll payment types, toll plaza driver behavior, and tidal lane management at Sydney Harbor. Although the API has been demonstrated to be feasible for enhancing Paramics capability, very little work has been done with other microscopic models, and no comparison study has been made among APIs in different simulation programs. This paper presents a user-defined API module for implementing a novel adaptive signal control technique, namely, a dynamic programming (DP) algorithm [15], [16] in AIMSUN, and eval- uating its enhanced capabilities via microscopic simulation. A dynamic link between the algorithm and the simulator is enabled through the API. By means of the link, the algorithm gets access to internal traffic data of the transportation system such as actual traffic demand, traffic queue, and vehicle speed. The algorithm then optimizes a signal control plan that is sub- sequently fed to and implemented in the transportation system to control the signal operations in running time. 1524-9050/$25.00 © 2008 IEEE