CLR Analytics


Papers on Vehicle Signature Technology

Shin-Ting Jeng and Lianyu Chu (2013) Vehicle Reidentification with the Inductive Loop Signature TechnologyIn Journal of the Eastern Asia Society for Transportation Studies, Vol. 10, 2013, pp. 1896-1915.

Abstract— This study proposed a real-time inductive loop signature based vehicle reidentification approach, RTREID-2M, which improved the previously developed RTREID-2 algorithm on two aspects: (1) Develop a cubic spline data imputation approach to replace the existing linear data imputation approach in order to improve the raw signature data quality; (2) Improve the time window setting for vehicles on High-Occupancy Vehicle (HOV) lanes so that vehicles on HOV lanes traveling with free flow speeds would be considered when generating candidate vehicle sets even during congestion time periods. In addition, a stratified-random sampling method was developed to effectively perform grouth-truthing task for evaluating the performance of the proposed RTREID-2M. The evaluation results showed desired performance for vehicle reidentification and travel time estimation under both free-flow and congested flow traffic conditions. The future research will focus on the potential applications and arterial vehicle reidentification utilizing the inductive loop signature technologies.

Jeng, S.-T., L. Chu, and S. Hernandez (2013). “Wavelet-k Nearest Neighbor Vehicle Classification Approach with Inductive Loop Signatures.”  Journal of the Transportation Research Board, No. 2380, pp. 72-80.

Abstract— In this study, a new vehicle classification algorithm was developed using the inductive loop signature technology.  There were two steps of the proposed algorithm.  The first step was to utilize the Haar wavelet to transform and reconstruct inductive vehicle signatures, and the second step was to group vehicles into Federal Highway Administration (FHWA) vehicle types using the K-Nearest Neighbor (KNN) approach with Euclidean distance classifier.  To determine the proper proportion of the wavelet to be applied for reconstruction and feature extraction, transformed signatures were examined with different percentages of large components of their corresponding wavelets.  To implement KNN, a library of vehicle signature templates for each FHWA vehicle class was composed.  The proposed vehicle classification algorithm demonstrated promising classification results with a 92.4% overall accuracy, and it can be applied to the real-world without the concerns of re-calibration and transferability using signature data from single loops.  Two additional vehicle classification schemes were applied for performance evaluation.  For the ISPE (Inductive Signature Performance Evaluation) classification scheme, which aimed to facilitate emission analysis and easy interpretation, the overall accuracy was 94.1%.  For the axle-based vehicle classification scheme proposed in this project, which aimed to group vehicles via usage and the number of axles of the vehicles, the overall accuracy was 93.8%.  Future research will focus on refining the signature template library for each FHWA vehicle type to further improve the performance of the proposed vehicle classification algorithm.  The selection of the value of k for the KNN approach will also be investigated.

Jeng, S.-T. and L. Chu.  Tracking Heavy Vehicles Based on Weigh-In-Motion and Inductive Loop Signature Technologies.  IEEE Transactions on Intelligent Transportation Systems, Vol. PP, Issue: 99, 2014, pp. 1-10, DOI: 10.1109/TITS.2014.2333003.

Abstract— Weigh-In-Motion (WIM) has been employed as a major technology to collect heavy vehicles’ data on the freeways. Because WIM is one of the most costly and sophisticated data collection system, how to effectively utilize the valuable WIM data and monitor WIM stations’ performance are especially important. In this study, we proposed an innovative and yet practical approach for heavy vehicle tracking that combines the use of both WIM data and the inductive loop signature data. The proposed multi-level vehicle reidentification approach was able to generate promising tracking performance with both inductive loop signatures and WIM data applied.

Shin-Ting Jeng and Lianyu Chu (2014) A High-Definition Traffic Performance Monitoring System with the Inductive Loop Detector Signature Technology, submitted to IEEE ITS Conference in Qingdao, China, Oct 8-11 2014.

Abstract— With continuing emphasis on transportation sustainability and fiscal stewardship, utilizing existing loop detector infrastructure to obtain more accurate, reliable, and comprehensive traffic system performance measures is desired by many transportation agencies. We found that the capability of the Inductive Loop Detector (ILD) signature technology to reidentify and classify vehicles along a section of roadway have the potential to provide better performance measures. Therefore, we proposed a high-definition traffic performance monitoring system (Traffic Monitor HD) based on the ILD signature technology and existing loop infrastructure for both freeway and arterial applications. Compared to the traditional performance measurement system, the advantages of the ILD signature technology allow Traffic Monitor HD to provide more comprehensive and accurate performance measurements, including point-based measures (i.e., vehicle counts, classification, and alerts on problematic detectors), section-based measures (i.e., travel time, speed, and estimates on emission), and O-D based measures (i.e., O-D matrix and trip travel time).

Lianyu Chu and Shin-Ting Jeng (2015) Field Testing of an Advanced Inductive Loop Detector Cards for Quality Inductive Loop Signatures, submitted to TRB 2015, January 11-15, 2015, Washington, D.C.

Abstract— Inductive Loop Signature Technology is a newly established approach for traffic performance surveillance, operational improvement, and demand management. The capability of the inductive loop signature technology to reidentify and classify vehicles has been demonstrated in a FHWA Small Business Innovation Research (SBIR) Phase I project. The ongoing Phase II study will develop a high-resolution traffic performance monitoring and analysis system (Traffic Monitor HD) based on the technology and existing loop infrastructure for both freeway and arterial applications. This paper presents a comprehensive field test of a new Inductive Loop Signature (ILS) card developed for the system, including the validation of the hardware compatibility, verification of the capability of the ILS card, and the investigation of the quality of the vehicle signature data. The testing results show that the ILS card is compatible with the 170 controller cabinet, the existing loops installed in the field, and the existing traffic data collection system. Based on the examination of the vehicle volume, reasonable waveforms of the vehicle signature by FHWA vehicle type, reliability of the communication channels, record timestamps, and signature data quality, it was found that USB mode is better than BT mode regarding data communication reliability. Live mode signature data output generally provides better signature data quality than the Complete mode and is preferred for real world implementation. This field test study concludes that the ILS card offers acceptable the signature data and thus it now ready to be further deployed.

Shin-Ting Jeng, Lianyu Chu,  Mecit Cetin (2015) Weigh-In-Motion Station Monitoring and Calibration using Inductive Loop Signature Technology, submitted to TRB 2015, January 11-15, 2015, Washington, D.C.

Abstract— Despite heavy vehicles representing a small portion of vehicles on the roads, they have significant influences on pavement, safety, environment, energy consumption, and the performance of traffic system. Weigh-In-Motion (WIM) is the major technology employed to collect truck data on the freeways over three decades.  However, WIM stations usually are not calibrated in a timely fashion and the calibration is mainly performed using five-axle single-trailer trucks once every half a year to three years.  A potential solution is to adopt a comprehensives remote calibration monitoring system.  Therefore, this study proposed an inductive loop signature-WIM based approach, which utilized both inductive loop signatures and WIM data to track heavy vehicles at WIM stations and generated “Matched Vehicle Pairs (MVPs)” for WIM station monitoring and calibration.  The algorithm was established based on a previously developed truck tracking algorithm, RTREID-2MT, and integrated with a Bayesian reidentification model to filter out the MVPs that were most likely incorrectly matched by the system.  The MVPs were then utilized for WIM station monitoring and temporary approximate calibration applications.  Case study showed that the upstream station reported low weights, while the downstream station reported high axle spacings.  The average offsets of the drive tandem axle spacing, Gross Vehicle Weight (GVW), and steer axle weight between the stations were thus derived from MVPs on a per lane basis and successfully applied to calibrate the problematic stations.