Traffic Data Anomalies in ATSPMs: History
View Latest Version
Please note this is an old version of this entry, which may differ significantly from the current revision.
Contributor:
Bangyu Wang
,
Grant G. Schultz
,
Gregory S. Macfarlane
,
Dennis L. Eggett
,
Matthew C. Davis

Automated traffic signal performance measures (ATSPMs) are an innovative technology that has garnered increasing attention in recent years due to their ability to collect and evaluate real-time and historical data at signalized intersections.

  • ATSPM
  • big data
  • moving average and standard deviation
  • performance measures

1. Introduction

Automated traffic signal performance measures (ATSPMs) are an innovative technology that has garnered increasing attention in recent years due to their ability to collect and evaluate real-time and historical data at signalized intersections. This technology enables traffic engineers and planners to optimize mobility, manage traffic signal timing and maintenance, reduce congestion, save fuel costs, and improve safety through data that are passively collected 24 h a day, 7 days a week [1][2][3][4][5].
The Utah Department of Transportation (UDOT) is one agency in the United States that has actively engaged in collecting traffic data and evaluating traffic signal performance throughout the state. UDOT utilizes ATSPM data to evaluate the quality of traffic progression along corridors while displaying unused green time that may be available from opposing traffic movements. The information derived from ATSPM data helps to inform UDOT (and other government agencies using ATSPM data) of vehicle and pedestrian detector malfunctions while measuring vehicle delay and recording vehicle volume, speed, and travel time [6][7]. A 2020 Federal Highway Administration publication estimated that UDOT’s ATSPM system saved taxpayers over USD 107 million in a decade by reducing traffic delays [8]. Additionally, an increasing number of states are either fully adopting ATSPMs or preparing to integrate them into their state transportation systems. Many state agencies leverage ATSPMs to support their daily operations and maintenance, guide transportation policy making, and shape future transportation master planning efforts.
In 2020, a Brigham Young University (BYU) transportation research team developed an analysis procedure that establishes meaningful thresholds for various signal system performance measures and created a statistical scoring tool and interactive data visualization tool to evaluate intersections across a collection of individual performance measures. This scoring tool and interactive data visualization tool were used to provide context to the historic quality of signal system operations across the state of Utah [9]. In the process of developing the scoring and data visualization tools, the research team found several inconsistencies or anomalies in the data, such as data switching, data shifting, missing data, and irregular curves. These anomalies present in the ATSPM data were evaluated, and it was determined that they have the potential to produce inaccurate results for any ATSPM data analysis [10]. Transportation agencies that use ATSPM data for policy making and performance analysis without understanding and correcting for data anomalies may misunderstand traffic conditions and misallocate resources.

2. Traffic Data Anomalies in Automated Traffic Signal Performance Measures

Over the past several years, a BYU research team has been evaluating the quality of signal operations using ATSPM data in Utah. The primary objective of this research was to evaluate performance measurement data collected through the UDOT ATSPM database. First, suitable performance measures were identified to evaluate traffic signal maintenance and operations by determining which performance measures had appropriate data available for analysis. Next, threshold values for each selected performance measure were established. Finally, a process for overall evaluation of the historic quality of signal systems statewide was developed. The research findings are documented in the literature [9].
During the previous research, several limitations and challenges were encountered. The most significant challenge arose from the incompleteness of the ATSPM database, with missing or extreme values for various performance measures across multiple signals. This presented unexpected difficulties in data evaluation and analysis, as the researchers had to contend with incomplete data. The limited availability of usable data also posed constraints in constructing the scoring method [10].
Various factors may be responsible for causing the data anomalies in the ATSPM datasets. First, Chang et al. [11] found that properly installing microwave sensors to detect traffic volumes and speeds is paramount to data accuracy, while the improper installation of detectors at intersections may lead to inaccurate data. Second, the ATSPM controller event log could be inaccurate when compared with the aggregated performance measure data at various timestamps. Any discrepancies present between the raw data and the aggregated data may appear. Third, the methodology and coding for each performance measure may have errors in the calculations and aggregation performed [12].
Data anomalies and outliers in the ATSPM datasets refer to individual data points that do not behave the same as data points adjacent to the time series. These points are generally isolated occurrences, but when multiple consecutive points behave in this manner, the pattern is called a subsequent outlier. Blázquez-García et al. [13] provided a review of methods used to detect these types of data anomalies in univariate and multivariate time series. Although various methods exist for addressing subsequent outliers, many rely on assumptions that are not present in the ATSPM data. For example, one of the most used methods to detect subsequent outliers in time series is called the HOT-SAX algorithm [14][15][16]. The algorithm was developed primarily to identify discords or the most unusual sequence in a time series, but the algorithm requires periodic data to discern an anomaly. Senin et al. [17] used the SAX algorithm, which can locate variable width subsequence discords, but this still relies on repeating data patterns to discern an anomaly. The ATSPM data provide non-periodic data with non-repeating patterns where the anomalies are present as discontinuous patterns. In these cases, methods used primarily for point outliers cannot be applied to detect subsequent outliers.
Multiple research projects have been conducted to visualize real data and compare data anomalies. The Georgia Department of Transportation (GDOT) SigOps metrics dashboard is one such tool that was created by GDOT to visualize ATSPM data longitudinally. Monthly, quarterly, and all-time summary tables are available with trending plots to compare changes in traffic-related performance measures over time [18]. The UDOT Watchdog report is a tool used by UDOT to evaluate signal controller data. The Watchdog report is an automatically generated summary of data errors found in traffic signal controllers over the previous 24 h, which are sent as an email to UDOT traffic engineers each weekday morning. These reports alert UDOT traffic engineers of detector issues that are manifested through inaccurate data [10]. Chamberlin and Fayyaz [19] compared continuous count stations (CCSs) data in Utah to evaluate turning movement count accuracy in ATSPM datasets. They found that the ATSPM volume data had many data anomalies present, mainly in the form of jump discontinuities, where volume data shifted up multiple hundred units for a time and then shifted back down to regular levels [19][20]. These research projects and applications can show when and where the data anomalies appeared but still need traffic engineering judgment to identify the data anomalies.

This entry is adapted from the peer-reviewed paper 10.3390/futuretransp3040064

References

  1. Bullock, D.M.; Clayton, R.; MaSckey, J.; Misgen, S.; Stevens, A.L. Helping Traffic Engineers Manage Data to Make Better Decisions: Automated Traffic Signal Performance Measures. ITE J. 2014, 84, 33–39.
  2. Lattimer, C.R.; America, A.N. Automated Traffic Signals Performance Measures; Publication FWHA-HOP-20-002; U.S. Department of Transportation, Federal Highway Administration: Washington, DC, USA, 2020.
  3. Smaglik, E.J.; Sharma, A.; Bullock, D.M.; Sturdevant, J.R.; Duncan, G. Event-Based Data Collection for Generating Actuated Controller Performance Measures. Transp. Res. Rec. J. Transp. Res. Board 2007, 2035, 97–106.
  4. Day, C.M.; Bullock, D.M.; Li, H.; Remias, S.M.; Hainen, A.M.; Freije, R.S.; Stevens, A.L.; Sturdevant, J.R.; Brennan, T.M. Performance Measures for Traffic Signal Systems: An Outcome-Oriented Approach; Purdue University: West Lafayette, IN, USA, 2014.
  5. Wu, X.; Liu, H.X. Using High-Resolution Event-Based Data for Traffic Modeling and Control: An Overview. Transp. Res. Part C: Emerg. Technol. 2014, 42, 28–43.
  6. Utah Department of Transportation (UDOT). ATSPM Frequently Asked Questions. Available online: https://udottraffic.utah.gov/ATSPM/FAQs/Display (accessed on 31 January 2022).
  7. Georgia Department of Transportation (GDOT). Automated Traffic Signal Performance Measures Component Details. Available online: https://traffic.dot.ga.gov/ATSPM/Images/ATSPM_Component_Details.pdf (accessed on 31 August 2022).
  8. Day, C.M.; O’Brien, P.; Stevanovic, A.; Hale, D.; Matout, N. A Methodology and Case Study: Evaluating the Benefits and Costs of Implementing Automated Traffic Signal Performance; Publication FHWA-HOP-20-003; FHWA and U.S. Department of Transportation, Federal Highway Administration: Washington, DC, USA, 2020.
  9. Wang, B.; Schultz, G.G.; Macfarlane, G.S.; McCuen, S. Evaluating Signal Systems Using Automated Traffic Signal Performance Measures. Future Transp. 2022, 2, 659–674.
  10. Schultz, G.G.; Macfarlane, G.S.; Wang, B.; Davis, M.C. Detecting Traffic Data Anomalies in Longitudinal Signal Performance Measures; Report No. UT-22.21; Utah Department of Transportation, Research and Innovation: Salt Lake City, UT, USA, 2022.
  11. Chang, D.K.; Saito, M.; Schultz, G.G.; Eggett, D.L. How Accurate Are Turning Volume Counts Collected by Microwave Sensors? In International Conference on Transportation and Development 2016; American Society of Civil Engineers: Reston, VA, USA, 2016; pp. 945–956.
  12. Georgia Department of Transportation (GDOT). Automated Traffic Signal Performance Measures Reporting Details. Available online: https://traffic.dot.ga.gov/ATSPM/Images/ATSPM_Reporting_Details.pdf (accessed on 31 August 2022).
  13. Blázquez-García, A.; Conde, A.; Mori, U.; Lozano, J.A. A Review on Outlier/Anomaly Detection in Time Series Data. ACM Comput. Surv. (CSUR) 2021, 54, 1–33.
  14. Keogh, E.; Lin, J.; Fu, A. Hot Sax: Efficiently Finding the Most Unusual Time Series Subsequence. In Proceedings of the Fifth IEEE International Conference on Data Mining (ICDM’05), Houston, TX, USA, 27–30 November 2005; IEEE: Piscataway, NJ, USA, 2005; p. 8.
  15. Keogh, E.; Lin, J.; Lee, S.H.; Herle, H.V. Finding the Most Unusual Time Series Subsequence: Algorithms and Applications. Knowl. Inf. Syst. 2007, 11, 1–27.
  16. Lin, J.; Keogh, E.; Fu, A.; Van Herle, H. Approximations to Magic: Finding Unusual Medical Time Series. In Proceedings of the 18th IEEE Symposium on Computer-Based Medical Systems (CBMS’05), Dublin, Ireland, 23–24 June 2005; IEEE: Piscataway, NJ, USA, 2005; pp. 329–334.
  17. Senin, P.; Lin, J.; Wang, X.; Oates, T.; Gandhi, S.; Boedihardjo, A.P.; Chen, C.; Frankenstein, S. Time Series Anomaly Discovery with Grammar-Based Compression. In Proceedings of the Edbt 2015, Brussels, Belgium, 23–27 March 2015; pp. 481–492.
  18. Georgia Department of Transportation (GDOT). SigOpsMetrics. Available online: http://sigopsmetrics.com/main/ (accessed on 7 July 2022).
  19. Chamberlin, R.; Fayyaz, K. Using ATSPM Data for Traffic Data Analytics; Report No. UT-19.22; Utah Department of Transportation, Research and Innovation Division: Salt Lake City, UT, USA, 2019.
  20. Utah Department of Transportation (UDOT). Continuous Count Station. Available online: https://data-uplan.opendata.arcgis.com/datasets/uplan::continuous-count-stations (accessed on 12 May 2022).
More
© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.
This entry is offline, you can click here to edit this entry!
ScholarVision Creations
Feedback