Search the USF Web site Site Map USF home page

Home

Center Identification Number:  77803   

Project Title:  Guidebook on Using Automatic Passenger Counters for NTD Reporting

Principal Investigator:

Xuehao Chu, Senior Research Associate
Phone: 813-974-9831
E-mail: chu@cutr.usf.edu

Institution:                             

Center for Urban Transportation Research
University of South Florida
Fax: 813-974-5168

External Project Contact:     

Diane Quigley
Florida Department of Transportation
(850) 576-2788
Email: diane.quigley@dot.state.fl.us

 

I.  Start and End Dates

Start Date: August 22, 2008                Expected End Date: June 1, 2010

I.  Project Objective/Problem Statement

Using electronic infrared beams or mechanical treadle mats, automatic passenger counters (APCs) have the ability to count transit passengers as they board and alight transit vehicles at individual stops.  When coupled with stop location information, archived APC data can be post-processed to generate disaggregate data both in time and space well suited for service planning.  An increasing number of transit agencies are considering APCs throughout the U.S.  Besides those agencies that have already implemented APCs in Florida, five others are testing and four agencies have plans to acquire them. 

While APCs are mainly used to gather data for service and operations planning, they can also be used to save data-collection costs for reporting to the National Transit Database (NTD).  To be eligible for the Federal Urbanized Area Formula Grant Program, transit agencies are required to report annual data on passenger miles (PMT) and unlinked passenger trips (UPT) to the NTD for each mode and service type (purchased versus directly operated).  Transit agencies can either report a 100-percent count or a sample-based estimate of each quantity.  When sampling, agencies had been allowed to use NTD-approved sampling techniques or any alternative sampling techniques (either APC-based or non APC-based) certified by qualified statisticians.  Unlike UPT, PMT is almost always a sample-based estimate for most services, and collecting sample data for estimating PM involves significant costs because information on both boarding and alighting activities at each stop is required for the sampled bus trips.  By rotating APC buses to cover pre-sampled bus trips that meet NTD’s minimal statistical requirements, agencies can significantly reduce or even eliminate manual collections and related costs.

Hope for such cost savings will be dramatically reduced for most agencies or may even disappear for many agencies under NTD’s new requirements for agencies using APCs for collecting NTD data.  One part of these requirements is an annual validation of APC data against manual methods using a random sample of at least 100 bus trips.  This sample-size requirement for annual validation dramatically reduces the expected cost savings from using APCs in collecting NTD data.  Moreover, the required 100 bus trips are likely to be similar or greater than what many agencies would need with manual ride-checks, and as a result there is little or no cost savings from using APCs.  Pinellas Suncoast Transit Authority (PSTA) in St. Petersburg, Florida is one example.  On April 25, 2007, its Board of Directors approved the purchase and installation of 10 APCs in addition to the 15 already installed.  Using an alternative sampling technique, however, PSTA collects fewer than 100 one-way bus trips each year for its directly operated bus services and still meets NTD’s minimal statistical requirements.  Overall, these new requirements put significant technical and financial burdens on transit agencies that would like to use APCs for NTD data collection.  Failure to meet these requirements means the loss of 5307 apportionment.  To avoid such loss, these agencies are likely to continue to use their current costly manual methods. 

II.  Objectives/Tasks

This proposed research focuses on using APCs to collect sample data for estimating annual passenger miles for fixed-route bus services.  One objective is to help transit agencies to avoid or reduce data errors by gaining a better understanding of potential errors from working with APCs and from post-processing archived APC data.   The second objective is to develop guidance for transit agencies to determine APC-based sampling plans.  The third objective is to develop guidance for transit agencies to meet the new NTD requirements at lowest costs possible to them.  The last objective is to explore how these requirements might be altered so that they would still achieve NTD’s original objectives, but put fewer burdens on transit agencies.

Task 1:  Project Management

This task will deal with on-going management activities of the project, preparing progress reports, and reviewing and editing final deliverables.

Task 2:  Understanding Potential Errors

This task will develop guidance to help transit agencies develop a better understanding of potential errors that commonly result from working with APCs and post-processing archived APC data.  This task will focus on non-random errors that lead to systematic under- or over-statement of passenger miles.  Besides raw counting errors by APCs, systematic errors can result from associating raw count data for individual stops to wrong stop locations and from attributing raw count data for individual bus trips to wrong bus trips.  In addition, errors in passenger miles can be substantially greater than these errors in counts of boarding and alighting.  Such reduced accuracy in passenger miles result because passenger miles are post-processed as the distance weighted sum of segment passenger loads; passenger loads are post-processed from the raw counts of boarding and alighting at individual stops; and the way passenger loads are determined allows errors in raw counts of boarding and alighting to accumulate.

Task 3: Sampling with APCs

This task will develop guidance to help transit agencies determine APC-based alternative sampling techniques in order to deal with random errors.  For most transit agencies, sampling requirements for NTD passenger-miles reporting are considerably less demanding than are other uses of the data such as monitoring loads or boardings by route, because the NTD precision requirement is only applied to a whole year’s sample aggregated systemwide. Therefore, meeting the NTD requirement should be easy for almost any transit system with APCs.  However, the NTD requires that alternative sampling techniques be statistically justified, and as a result, more rigor is needed in determining an APC-based sampling technique for NTD reporting purposes.

Task 4: Meeting APC Requirements

This task will develop guidance for transit agencies to meet NTD’s new requirements for using APCs to collect NTD data at the lowest cost possible to agencies.  Effective with the 2005 Report Year, the requirement is that FTA approves the APC methodology, the implementation of a new APC system, and the APC maintenance and benchmarking plan for each transit agency.  Failure to obtain prior FTA approval for Report Year 2005 and future report years will result in APC-derived passenger mile data not being included in the Urbanized Area Formula Program apportionment.  The specific requirements are presented in the following box. 

 

 

Text Box: For the first year: 
 §  The APC, UPT and PMT data are validated against a separate data sample covering a full year
§  Data sampling meets FTA requirements for 95% confidence and 10% precision using: 
o    A random sampling of vehicle trips
o    Assigning the buses with an APC to the vehicle trip sampling plan. 
 In subsequent years: 
 §  Calibration of the APC equipment every year using a random sample of at least 100 bus vehicle trips using ride checkers to collect the UPT and PM data. Cameras do not replace ride checkers. The APC equipment needs to be randomly distributed by route, by day, and time of day to avoid sampling bias that may result from placement of APCs on heavily traveled routes in peak times, and in times of seasonal peaks. 
§  The transit agency submits to FTA documentation of the UPT and PMT data collected by ride checkers compared to APC derived UPT and PMT data and the statistical variance between the two data sets. 
If a transit system uses APCs for both directly operated and purchased transportation bus services, separate samples of at least 100 bus vehicle trips are required for each type of service.
 
 

 

 

 

 

 

 

 

 

 

 

 

 

Task 5:  Reconsidering Requirements

This task will explore how these new requirements on using APCs for collecting NTD data might be altered so that they would still achieve NTD’s original objectives, but put fewer burdens on transit agencies.  At a minimum, for example, the minimum sample size of 100 bus vehicle trips for annual validation should be proportional (e.g., ⅓, ¼, etc.) to the pre-APC annual sample size that meets FTA’s 95% confidence and 10% precision levels.

Task 6:  Preparing Guidebook

This task will develop a guidebook to transit agencies that use APCs to collect NTD data.    Specifically, it will cover three related areas: 1) guidance to help transit agencies develop a better understanding of potential errors that commonly result from working with APCs and post-processing archived APC data; 2) guidance to help transit agencies determine APC-based alternative sampling techniques in order to deal with random errors; and 3) guidance for transit agencies to meet NTD’s new requirements for using APCs to collect NTD data at the lowest cost possible to agencies.  In addition, slides will be prepared for possible inclusion of the developed guidance in the annual NTD training in Florida.

 

III. Deliverables

Project deliverables will include quarterly progress reports, a draft guidebook, a final guidebook, and slides for the annual NTD training for Florida agencies.

Progress Reports

Progress reports will be submitted on a quarterly basis to the Research Center for processing.  The first progress report will become due 90 days after a Notice to Proceed is issued to the consultant by the Department.  They may be sent in MS Word format to Sandra Bell, Contracts Administrator at Sandra.Bell@dot.state.fl.us.

Progress reports must include the following information:

1.      Contract Number, Work Order Number and Title

2.      Work performed during the month

3.      Work to be performed in the following month

4.      Requested modifications (i.e., to funding, schedule, or scope) *see separate instructions for requesting approval of modifications

5.      An updated Exhibit “D” Progress Schedule.

Failure to submit progress reports in a timely manner may result in termination of this task work order.

Draft Guidebook

The draft guidebook will be submitted to Sandra Bell at sandra.bell@dot.state.fl.us .  Draft reports must be prepared in accordance with the Guidelines for Preparing Draft Final and Final Reports.

Final Guidebook

A total of 13 hard copies of the final guidebook and 8 copies of the final report in MS Word on CD and one unbound original will be delivered no later than the end date of the task work order.

Project Certification 

The Sponsored Research Office or appropriate authority will submit as a final deliverable a project certification prepared according to university compliance standards.

IV.  Project Schedule

Progress Schedule

It is anticipated that the project will be completed within 18 months of the written notice to proceed from the Department. 

V.  Project Budget

VI. Equipment

No need for non-standard equipment is expected for conducting the proposed research.

VII. Travel

No travel is expected to achieve the project objectives.

VIII. Student Involvement

It is expected that this project will use a graduate student who is yet to be recruited and who will also work on other research projects.  The graduate student will be recruited among those with experience of processing huge datasets with Microsoft SQL Server.  For this project, this graduate student will mainly work on processing archived APC data sets using Microsoft SQL Server.  The hourly rate used in the budget includes in-state tuition costs.  Including tuition in the hourly rate makes sure that tuition will only be covered for the hours in which the student works on the project.

IX. Relationship to Other Research Projects

IX. Potential Benefits of the Project

This proposed research focuses on using APCs to collect sample data for estimating annual passenger miles for fixed-route bus services.  One objective is to help transit agencies to avoid or reduce data errors by gaining a better understanding of potential errors from working with APCs and from post-processing archived APC data.   The second objective is to develop guidance for transit agencies to determine APC-based sampling plans.  The third objective is to develop guidance for transit agencies to meet the new NTD requirements at lowest costs possible to them.  The last objective is to explore how these requirements might be altered so that they would still achieve NTD’s original objectives, but put fewer burdens on transit agencies. 

 

National Center for Transit Research · at the Center For Urban Transportation Research · University of South Florida · 4202 E. Fowler Ave., CUT100 · Tampa, FL 33620-5375 · (813) 974-3120 · (813) 974-5168 · www.nctr.usf.edu · Comments: webmaster@cutr.eng.usf.edu