Transit Service Reliability: Analyzing Automatic Vehicle Location (AVL) Data For On-Time Performance and to Identify Conditions Leading to Service Degradation

(Center Identification Number: 79050-02-B)

Principal Investigators:

Fabian Cevallos, Ph.D., Transit Program Director
Lehman Center for Transportation Research (LCTR)
Florida International University (FIU)
10555 West Flagler Street, EC 3609
Miami, FL 33174
(305) 348-3144

Sean Barbeau, Director, Research Associate
Center for Urban Transportation Research (CUTR)
University of South Florida (USF)
4202 E. Fowler Avenue, CUT 100
Tampa, Florida  33620-5375
(813) 974-7208


Automatic Vehicle Location (AVL) systems are computer-based vehicle tracking systems that function by measuring the real-time position of each vehicle and relaying this information back to a central location. AVL systems are most frequently used for fleet management to identify the location of vehicles for a variety of purposes including: improved dispatch, operation efficiency, and faster response times to disruptions in service, such as vehicle failure or unexpected congestion; quicker responses to threats of criminal activity; and improved data for future planning purposes.

An AVL system tracks vehicles using one of the following location technologies: GPS, signpost and odometer interpolation, ground-based radio (Loran-C), or dead reckoning. GPS is the most effective tracking system for transit because it is the most accurate, can be used with variable routing and scheduling, and does not require purchase, installation, or maintenance of wayside equipment. The GPS system works through a network of orbiting satellites that transmit signals to the ground. Special receivers on each vehicle read the available signals and triangulate to determine their position. The geographic location along with the date, time, and other operational data are then sent to the transit agency. Data are transmitted, at established polling intervals, to the transit center with use of radio or cellular communications and can be used immediately for daily operations or archived for further analysis.

Transit agencies are increasingly collecting and analyzing AVL data to support a variety of operations, scheduling, and service planning activities. Data can be used as source for key performance indicators, responding to service complaints, or reviewing and improving the quality of service. The use of historical data can help identify problems that have occurred in the past. Further, analysts can identify recurring problems and develop solutions to these problems.

Data from an AVL system can also be used to measure, monitor, and improve service reliability, also known as on-time performance. Note that adherence to the schedule is a matter of service quality to transit passengers. From the service provider perspective, schedule adherence reflects the quality of the service plan: schedules and operations control. It is also important to clarify the difference between schedule adherence and on-time performance. Schedule adherence refers to the difference between real time and scheduled times of arrival or departures times, usually presented in minutes. On-time performance (OTP), on the other hand, is a percentage value used to indicate buses arriving or departing late, on-time, or early. Depending on the different technologies, on-time performance can be calculated using arrival, departures, or possibly a combination of both.

As on-time performance is very important to the transit customer, OTP strategies can help improve customer satisfaction and attract new transit riders. Reliability is one of the areas that transit agencies can use to improve service at a relatively low cost. Therefore, the benefit to cost ratio of improving on-time performance is expected to be significant.

To improve transit service reliability, there is also a need of a systematic review of historical AVL data in order to identify recurring service problems and to see if there are conditions in the data that exist in the time period preceding the service problem. Early identification of these conditions can help transit agencies make intelligent decisions and determine the best course of action for avoiding service degradation. This can help enhance quality of service and customer satisfaction.

To assist in this effort, the Lehman Center for Transportation Research (LCTR) at Florida International University (FIU) in collaboration with the Center for Urban Transportation Research (CUTR) at the University of South Florida (USF) propose the development of a methodology and framework for improving transit service reliability, which can be used towards the implementation of a computerized system for the application of these concepts.

Project Objectives

The main objective of this scope of work is to conduct research on the use of AVL data for improving transit service reliability. This can be achieved by using better on-time performance techniques and by identifying conditions leading to service degradation that can assist transit agencies in providing higher quality of service.

This research will concentrate in two related areas for improving service reliability: 1) investigate the challenges and issues towards measuring, monitoring, and improving on-time performance and 2) identify service conditions observable in AVL data that precede service problems. Output products would include recommendations for improving on-time performance and a list of candidate factors or conditions that could lead to service degradation and how transit agencies could use this information. The results from this research may warrant the development of computerized tools as the next step.

To achieve this, the FIU/USF research team proposes a series of work tasks that include conducting the literature review, assessing the AVL system and the datasets, examining challenges and opportunities, creating a methodology, developing a framework, and producing the final report.

Work Tasks

Task 1: Literature Review

This research will start with a review of the literature on AVL systems and the use of data. It will include, but not be limited to, TCRP reports, TRB papers, related papers from other journals and conferences, and available reports from transit agencies, consultants, and vendors. This task will focus on identifying and reviewing the AVL technologies and current and innovative use of the data. The research team will also make use of the resources available at the website of the Florida Advanced Public Transportation Systems (APTS) Program:

Task 2: Assessment of AVL Systems and Datasets

The research team will compile information regarding the AVL systems and will identify the datasets needed for the development of a computerized system to improve service reliability.

Accessibility to real-time as well as archived AVL data will be investigated. AVL datasets, parameters, standards, and related elements will be assessed as part of this effort. Potential technical challenges will be documented and suggestions for improvement will be presented. A transit agency with AVL experience will be included as part of this task.

Task 3: Challenges and Opportunities

As transit agencies often rely on limited Planning, Operations and Information Technology (IT) staff to analyze the data, the need for sophisticated tools to assist agencies are critical. The purpose is to assist agencies with limited resources with improving service quality for customers. For this, it is important to understand the quality and the level of use of real-time and static AVL data within the transit agency, in particular in the areas of operations, planning, and scheduling. Additionally, issues like available resources, dealing with vendors, and the benefits of partnering with local universities need to be considered. Success stories will be included like the one from Tri-Met in Portland, Oregon, who has had an ongoing relationship with researchers at Portland State University.

While there are a number of research papers and a TCRP synthesis on ways to use AVL data, transit agencies generally analyze their own data to observe operations issues, such as schedule adherence and service quality. However, the research team is not aware of any research that uses AVL data to identify factors or conditions in the data that immediately precedes service quality deterioration at the route-level. Further, the research will cover areas that have not been fully addressed in previous research such as data quality, standards, and methodologies for measuring and monitoring on-time performance.

Task 4: Development of Methodology

Based on the datasets from the AVL system, methodologies for addressing on-time performance and for identifying conditions towards service degradation will be developed. It will consider temporal and spatial issues that impact service reliability.

For on-time performance, the following items will be addressed in the methodology: data quality, schedule adherence parameters, missing records, extreme values or outliers, end-of-line (EOL) and other issues that may have an impact on the way OTP can be measured, monitored, and managed. The development of the methodology is a key task in this project, as it will be the basis on how on-time performance will be calculated.

For identifying the conditions for service degradation, a flow chart, algorithm, or a step by step process will be investigated. Conditions such as the continued reduction of spacing among vehicles that can lead to bus bunching will be examined. Similarly, headway inconsistency and service gaps that departs from the schedule will be studied. In general, the research team will investigate factors that can impact the quality of service such as service frequency and headway maintenance throughout the day as well as other conditions related to particular route service characteristics.

Methods toward measuring anomalies before service disruptions will be developed. Relationships among service problems shown in the AVL data and preceding the service deterioration will be investigated. These relationships could be identified through statistical analysis or other methods. Candidate factors or conditions observable in AVL data that could lead to service degradation will be assessed and a measuring and/or a predicting methodology will be developed. The developed methodology can be used for effective monitoring and for decision making. These methodologies can be tested with data provided by the selected transit agency.

Task 5: Creation of a Framework

This task is to put together a framework for the development of a tool that can be used to enhance on-time performance and for reducing service disruptions. Developing a framework can facilitate the development process before the application is developed and deployed at the transit agencies. By using a computerized system, management and staff can help monitor, manage, and improve the efficiency of the transit system.

The framework will include the different components for measuring on-time performance and for identifying conditions towards service degradation. It will use AVL data for effective transit service management and enhanced delivery of services, considering route service characteristics.

With an emphasis on performance management, problems can be observed and improvements can be made. The framework will be the foundation of a system that can assist managers, supervisors, schedulers, and planners in their quest for providing quality transit service. The proposed framework can be used to develop a software application to measure, monitor, and help transit agencies make intelligent decisions. The system can also help operations with real-time control as well as to generate reports based on archived data for planning and scheduling purposes. The developed framework can be used as a model for the development of a working prototype that can be tested and implemented at transit agencies.

Task 6: Final Report

A draft final report documenting all aspects of this research will be prepared and submitted to the National Center for Transit Research (NCTR) for review and comments. This will include the literature review, assessment of the AVL system and the datasets, presentation of the challenges and opportunities, a description of the methodology, and the development of the framework. The final assessment and recommendations section will address how transit agencies can use this information and will identify future topics of research. Based on the feedback from NCTR, the report will be revised and finalized, and resubmitted.

Designated Personnel

  • Dr. Fabian Cevallos, Transit Program Director at LCTR, will serve as the Principal Investigator (PI) and will be responsible for this project. He will also be directly involved with the development of the methodology, framework, and all the technical aspects of the project.
  • Dr. Albert Gan at LCTR will serve as Co-Principal Investigator (Co-PI) and will provide overall technical assistance.
  • Sean Barbeau, Research Associate at CUTR, will assist with the development of the methodology and framework.
  • A Research Associate at LCTR will assist with the literature review, the creation of the methodology, and development of the framework.
  • A graduate student will be assigned to assist with this project.

Project Schedule

18 Months

Project Budget

Total Project Cost     $185,000

Leave a Reply

Please complete the simple math problem prior to submitting your comment. This reduces spam. Thanks! * Time limit is exhausted. Please reload CAPTCHA.

This site uses Akismet to reduce spam. Learn how your comment data is processed.