HomeMy WebLinkAbout29. Passenger Tracking System for GO Transit OfHKOJH
City of Oshkosh - Transportation Department ON THE WATER
926 Dempsey Trail, Oshkosh, WI 54902 (920) 232-5342 (920)232-5343 fax
MEMORANDUM
TO: Honorable Mayor and Members of the Common Council
FROM: Christopher Strong, P.E., Director of Transportation
DATE: November 1, 2012
RE: UPDATE ON PASSENGER TRACKING SYSTEM FOR GO TRANSIT
The first goal listed among five that the Common Council assigned to the City Manager at its
February 14, 2012 meeting is as follows:
"1. Develop a tracking system for the Oshkosh Transit System.
Plan the implementation of a system to track and provide data reporting and analysis of
all bus users by location. Include the implementation plan and required resources for the
tracking system in the 2013 budget. "
I presented some initial research into this goal at the May 9, 2012 Council meeting. Three options
were discussed:
• Full Electronic Fare Payment System
• Automatic Vehicle Location with Tablet Interface
• Automated Passenger Counting
It was noted at that time that none of these options perfectly satisfies the Council's goals for the
passenger tracking system, but they individually or collectively could provide valuable
information. The Council requested that staff identify transit services in similar sized cities that
have implemented one or more of these technologies, and learn more about their experiences
with these technologies. Council members were interested in knowing how these other transit
systems have used these technologies, whether they have been beneficial, what changes they
would make, and similar questions.
Electronic Fare Payment System Alternative
Instead of researching all three technologies, staff focused on automated passenger counting
systems. The full electronic fare payment system was the most expensive of the three options,
and was initially considered only in light of a grant application which GO Transit submitted to
the Federal Transit Administration. In the months since the May 9 presentation, staff learned that
we were unsuccessful in securing that grant. The cost of this system, along with the negative
City of Oshkosh—Department of Transportation 1
Update on Passenger Tracking System for GO Transit/November 1,2012
impacts it would have on existing fare and operations policies, recommend against further
consideration of this alternative.
More and more transit systems, including Wisconsin peer systems such as Green Bay Metro and
Valley Transit, are implementing electronic fare payment systems for a variety of reasons, such
as passenger convenience, revenue control, auditing, security and data. Therefore, we will
continue to pursue this in the future. However, due to the initial high cost and the lack of
supporting funding, this does not appear to be an ideal solution for the passenger data
requirements sought by the Council.
Automatic Vehicle Location(AVL)with Tablet Interface
The automatic vehicle location with tablet interface approach was set aside for a different reason.
While many systems have automatic vehicle location, we are not aware of any transit system
which is trying to integrate AVL with a manual fare collection method. Some systems have an
automatic vehicle location system which tracks the vehicle's progress through its route without
any information on fare collection or passenger counting; other systems have an AVL system
integrated with electronic fare collection. Our proposed concept falls between these two areas,
and we are not aware of any peer systems that have followed a similar approach.
Staff sees immediate value in this technology as it can provide real-time customer information,
improve reporting of passenger fare information, and can be used to count passenger boardings
by location. Therefore, staff has been working with a couple of vendors, a consultant and IT staff
since mid-2011 to pilot test this idea. Under the most advanced scenario that staff has considered,
this technology would be capable only of recording passenger boardings, not alightings. While
this would be better than the data currently available, it would not generate the extent of
information sought by the Council.
Automated Passenger Counting
Staff sent out a broadcast e-mail to Wisconsin transit systems regarding their experience with
automated passenger counting (APC). Staff followed up with three systems who volunteered
having experience with APC: Duluth(which also serves Superior in Wisconsin), Milwaukee
County and Madison. In addition to these case studies, staff identified a 2008 national research
report which provides a snapshot as to the state of the practice with APC nationally. The report
consists of a national survey of transit agencies regarding their experience with APC, and more
extensive case studies of six APC implementations (one of which also happened to be Madison).
This report is available at http://onlinepubs.trb.org/onlinepubs/tcrp/tcrp syn 77.pdf.
According to the report,transit agencies that use APC find continued value in it. Systems that
have had problems with it, as did Madison on an earlier iteration, plan to improve the technology
rather than discard it. The report finds, "APCs provide a rich ridership and travel time database at
a finer level of detail than farebox or manual counts, even for agencies with only a few APCs." It
also notes, "Transit agencies that have worked through the myriad issues associated with APC
implementation cannot imagine life without APCs."
City of Oshkosh—Department of Transportation 2
Update on Passenger Tracking System for GO Transit/November 1,2012
That said, there are a number of findings from the report that cause some concern:
• "An APC system does not work automatically." In other words, installing and using the
technology is not a simple "plug and play"type of solution. Some of the additional work
that is required for successful implementation includes matching data to stops, cleaning
and validating data, generating reports, hardware and software maintenance, and
integration with other data systems.
• "APC implementation is not simple, and the first year is the most difficult."Moreover,
the report notes elsewhere that APC implementation requires a de-bugging and
calibration period, which agencies report takes an average of 17 months. Most APC
customers (75 percent) indicate some problems with APC installation. This will make it
difficult to have good data during the early months of implementation of the new route
system.
• "Staffing presents a challenge, especially to small and medium-sized agencies. ...
Successful implementations are characterized by... a dedicated maintenance technician or
group of technicians who assumes primary responsibility for hardware issues. Agencies
may not have the staff available of may not have staff with the right mix of skills."
The report's case studies covered six transit systems: Ottawa [Canada], Denver, Buffalo, Reno,
Madison, and Portland. The smallest of these systems is Reno, with 62 buses in peak service. The
following are some highlights from some of the case studies:
• Ottawa has more than 25 years of generally positive experience with APC, but
interestingly has APC on less than 10 percent of their peak fleet (90 of 991 buses). They
advised it could take up to three years to fully implement APC.
• Denver has also had positive experience with APC. They implemented their APC over a
three-year phased implementation. They maintain that a robust APC program can be
implemented over only 20 percent of the fleet(which consists of 921 buses and 57 light
rail vehicles during peak service).
• Buffalo started implementing APC about 10 years ago, and now covers approximately 40
percent of their fleet(280 buses in peak service). They experienced many accuracy issues
with their deployment, and as such conclude that deployment of APC is an issue that
requires supervision.
• Reno bought its APC as a part of a broader technology package in 2002. Reno reports
moderate satisfaction with their system. While they like the stop-level detail regarding
boardings and alightings,they have had issues with data quality. They said that agencies
need lots of resources and knowledge in both hardware and software to make APC
succeed.
• Madison bought its first system in 2004 and has had problems with data errors; they
report being moderately dissatisfied in APC's ability to deliver useful data on passenger
counts. The report noted that staff has had limited opportunity to dedicate the time needed
to iron out the problems they have had. They are in the process of procuring new APC
systems.
• Portland, which has 526 buses and 81 light rail vehicles in peak service, has had a long
track record with APC, including 25 years of experience, with coverage on over 75
City of Oshkosh—Department of Transportation 3
Update on Passenger Tracking System for GO Transit/November 1,2012
percent of its fleet. As an early adopter, Portland's in-house IT staff developed numerous
database applications to analyze data based on their needs.
Cost Estimate
Based on an analysis of survey responses, the report calculated an average cost of$6,600 per
vehicle for purchase and installation of APC equipment, and another $600 per vehicle for
operations and maintenance (in 2008 dollars). The report urges caution in interpreting these
figures, as APC has often been procured in conjunction with a larger technology purchase, so its
costs may be difficult to isolate. Staff talked with one vendor who indicated that the per-vehicle
capital costs could be lower, depending on how heavily involved staff is in the installation and
initial testing. However, it is also possible that costs may be higher for a smaller implementation
of this technology.
Operations and maintenance costs are exclusive of staff labor time involved in validation,
integration, maintenance, and other activities. If GO Transit were to adopt APC, it is
recommended that most, if not all, of these activities be contracted out, which would be an
additional cost.
Recommendation
Staff continues to work on the pilot project to create a tablet interface, in conjunction with AVL,
to simplify collection and reporting of fare data. This is a potentially low-cost solution that can
provide significant benefit in improving staff productivity and traveler information. If staff is
successful in developing this application, this will create a large-scale, near real-time database of
passenger boarding information on all of our routes. It cannot capture alighting information, so
this is only part of the picture requested by Council. However,this application can help our
organization to develop processes to verify, use and analyze these data, which will be important
for any future investments in similar technology. In addition,this data can provide a very
effective way to validate the accuracy of boarding information collected through APC, should it
be implemented in the future.
APC implementation appears to be the best option to track both boardings and alightings. If
Council is interested in continuing to pursue this, there are at least a couple of possible options.
• Fleet-wide Adoption. The Council could fund full implementation of APC on all of our
buses in the 2013 (or a future year's) Capital Improvement Program.
• APC on Newer Vehicles. GO Transit could include APC as a requirement in new bus
procurements, and retrofit the four 2010 hybrid buses. This would provide coverage over
approximately one-third of our fleet, which supports regular sampling over our entire City
service.
The following table lists some of the advantages and disadvantages with each approach. The
advantages of one approach often show up as disadvantages of the other approach. In either case,
we should not expect usable data from this system until at least 2014.
City of Oshkosh—Department of Transportation 4
Update on Passenger Tracking System for GO Transit/November 1,2012
These recommendations relate only to City vehicles. GO Transit operates Route 10 under a
service contract where the vehicles are provided by a contractor. This contract includes language
which supports the AVL pilot test described earlier; however, it does not include language
related to the installation of APC on contractor vehicles.
Staff supports a smaller scale implementation of APC, along the lines of Option 2, as a way to
expand upon the AVL pilot test, and enhance our passenger data.
Advantages Disadvantages Cost Estimate
Option 1: • Allows for a standard • Retrofits can be more • Capital:
Fleet-wide implementation which complicated, especially $150,000
Adoption can simplify on older vehicles • Operations and
(17 buses) maintenance and • Larger number of Maintenance:
training requirements vehicles will compound $10,000 per year
• Does not introduce the potential number of
additional complexity in maintenance issues
assigning buses to • Will not cover Route 10
routes
Option 2: • Lower cost of • May not generate • Capital: $50,000
APC on implementation sufficient data to be • Operations and
Newer • Lower risk exposure in useful Maintenance:
Vehicles the event of • Requires additional $6,000 per year
(2010 and technological challenges work in assigning buses
2013 only; • Lower number of to get good cross-
6 buses) vehicles reduces time section of data
for calibration and • Standardization could
validation be difficult
• Smaller scale of
implementation may
discourage its use, and
make it seem less
important
• Will not cover Route 10
Respectfully submitted, Approved:
hristopher Strong, P.E. Mark A. Rohloff
Director of Transportation City Manager
City of Oshkosh—Department of Transportation 5