IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, VOL. 16, NO. 4, AUGUST 2015 1667 Monitoring the Speed, Configurations, and Weight of Vehicles Using an In-Situ Wireless Sensing Network Wenjing Xue, Dong Wang, and Linbing Wang Abstract—The Virginia Polytechnic Institute and State Univer- sity is developing an integrated transportation monitoring system that will be capable of monitoring pavement and traffic simulta- neously. As part of the system, a backcalculation method, which is presented in this paper, is used to estimate the speed, weight, and configuration of passing vehicles based on the pavement responses collected by in-situ pavement sensors. This method is still in its preliminary stage but could be very helpful in achieving real-time weigh-in-motion and traffic classification once completed in the future. A Gaussian model is used to describe the distribution of the horizontal strain induced by passing vehicles. The parame- ters of the Gaussian model are correlated with various loading conditions, including the weight and the configuration parameters of the passing vehicles, as proved by finite-element simulation and experimental measurements. The backcalculation process is efficient and valuable, considering the accuracy of the estima- tion with a low computational cost. The whole method is simple and straightforward and can be conveniently used in real-time monitoring. Index Terms—Signal processing algorithms, traffic control, weight measurement, wireless networks. I. I NTRODUCTION C OLLECTING reliable traffic flow information is impor- tant for intelligent transportation systems (ITS) to make appropriate decisions on traffic signal timing and optimization. Based on different purposes, the target information of traffic flow for a specific section might include speed, traffic volume, classification, and axle load. Many efforts have been devoted to improving the information collection of ITS with various technologies in recent years. In modern society, heavy traffic including congestion can be observed all over the world, which makes traffic volume and speed important for traffic management [1]. Conventional sensing technologies for traffic counting and speed measuring can be categorized into two groups: intrusive and nonintrusive sensors. Intrusive sensors are directly installed on/in pavement Manuscript received April 4, 2014; revised June 16, 2014 and August 29, 2014; accepted October 16, 2014. Date of publication November 11, 2014; date of current version July 31, 2015. The paper was submitted on March 12, 2014 for review. This work was supported by the Mid-Atlantic Universities Transportation Center. The Associate Editor for this paper was Q. Kong. W. Xue is with the Virginia Polytechnic Institute and State University, Blacksburg, VA 24060 USA (e-mail: wenjingx@vt.edu). D. Wang was with the Virginia Polytechnic Institute and State University, Blacksburg, VA 24060 USA. He is now with the WorleyParsons Group, Sydney, N.S.W. 2060, Australia (e-mail: wenjingx@vt.edu). L. Wang is with the Virginia Polytechnic Institute and State University, Blacksburg, VA 24060 USA (e-mail: wangl@vt.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.2014.2364186 surface layers and include inductive loops, magnetometers, mi- croloop probes, pneumatic road tubes, and piezoelectric cables [2]. Nonintrusive sensors, which are usually mounted above the lane of traffic, can be installed and maintained safely and with minimal disruption of traffic. The technologies normally used in nonintrusive sensors include infrared [3], microwave radar [4], Global Positioning System [5], wireless sensor network [6], acoustic [7], bluetooth sensors [8], and image processing [1], [9] technologies. Weigh-in-motion (WIM) has been developed to obtain the weight of a vehicle while the vehicle is moving. Since the con- cept was brought up 60 years ago [10], WIM technologies have been used increasingly around the world to control the weight of heavy vehicles and protect pavement and other transportation infrastructures [11]. There are several major types of sensors used for WIM stations: piezoelectric sensors, capacitive mats, bending plates, load cells, and optic fibers. Traffic classification is used to measure the configuration pa- rameters of passing vehicles and classify them for traffic anal- ysis and management. Traffic classification plays an important role in ITS for the analysis and management of traffic flow. Cur- rent vehicle classification technologies can be grouped into four major categories: axle based, vehicle length based, machine vision based [12], and magnetic signature based [13]. Popular sensing technologies used in vehicle detecting are traditional piezoelectric sensors, magnetic sensors [14], inductive loops [15], and fiber optic sensors [16]. In addition, some researchers also used strain gauges to detect and classify vehicles [17]. Today, the transportation management system is expected to be integrated and intelligent; hence, it is necessary and beneficial to combine different monitoring systems together. In this situation, it will be very profitable to combine the monitoring systems for traffic and pavements together due to the interdependence between traffic and pavements. In the last few decades, a wide variety of sensors have been developed for pavement monitoring purposes [18]–[20]. At the same time, pavement research facilities (test roads), such as MnRoad [21] and Virginia Smart Road [22], [23], have become an integral part of pavement research and engineering. With the development of pavement sensing technologies, traditional pavement monitoring systems have been combined with traffic information collection systems, such as WIMs and traffic classification systems, by some transportation agencies. For example, the Florida Department of Transportation (FDOT) has approximately 350 traffic classification and WIM sites lo- cated throughout the state, including thousands of piezoelectric sensors [24]. Because of the low survival rate of the piezo- electric sensors, the FDOT started the development of a fiber U.S. Government work not protected by U.S. copyright.