|
ITE Traffic Incident
Management E-mail Discussion Group Digest #21
January 23 to February 2, 2000
This digest contains the following
messages:
#1.
TIM: Temporary freeway detection/surveillance -Reply - from Dan Howard
#2. TIM: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY -Reply - from
Clifford Conklin
#3. RE: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY - from O'Laughlin,
John B.
#4. Re: TIM: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY -Reply - from Ken
Brooke
#5. TIM: RE: Video: Why Can't We Talk? -Reply -
from
Helman, David
#6. Re: TIM: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY -Reply - from Dick Raub
#7. RE: TIM: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY -Reply - from
O'Laughlin, John B.
#8. TIM: Traffic Incident Management Congestion Ratios
- from Corbin, John
#9. Re: TIM: Traffic Incident Management Congestion
Ratios - from Ken Brooke
#10. Re: TIM: RE: Portable Cameras - from
Larry Corcoran
#11. RE: TIM: RE: Portable Cameras - from
Bruce Churchill
#12. Re: Traffic Incident Detection - from
Ken Brooke
#13. TIM: Re: Traffic Incident Detection - from
Barry Drogin
#14. Re: TIM: Re: Traffic Incident Detection -
from Ken Brooke
#15. Re: TIM: Re: Traffic Incident Detection -
from Barry Drogin
Disclaimer: Topics
that do not pertain to Traffic Incident Management have been omitted from
the discussion group digests to save the reader time when reviewing
messages, and to save space on the web server. See
Discussion Group Etiquette for examples of
other reasons a message may have been omitted from a digest.
-------------------- 1 --------------------
#1. TIM: Temporary freeway detection/surveillance
-Reply - from Dan
Howard
Top
Date: Sun, 23 Jan 2000 11:41:42 -0500
From: Dan Howard <dhoward@gw.dot.state.ny.us>
Subject: TIM: Temporary freeway detection/surveillance -Reply
While not exactly answering your question, this is a response for a
similar condition.
Two years ago NYSDOT Region One (Albany) let a contract for
permanent traffic detection and surveillance (loops and cameras).
Four (future) loop stations were installed with Whelen Wide beam RADAR
detectors instead because the highway was scheduled to be reconstructed
the following two construction seasons.
This allowed us to collect average speeds from all sites, even the ones in
the work zones. Now that one of the subsequent projects is complete,
we are now obtaining traffic volumes from three of the sites that have
converted from RADAR to loops as part of the roadwork. Next year,
the fourth site will be converted.
Similarly.....
On a separate loop contract this past year we had the contractor install
PVC conduit under loops at one location. 3M will be loaning us some
of their Canoga microloops for comparison speed/volume sampling (we also
have a CCTV camera that will allow us to manually collect volumes as
well).
The hope is that the Canogas will work as well as the loops and in the
future, Canoga sites are located in work zones will still provide freeway
detection. Since they are installed in conduits under the roadbed,
the Canogas could even be moved laterally in conjunction with any lane
shifts associated with the road work.
Dan Howard
Operations Engineer
Capital Region Transportation Management Center
Albany, NY
<snip>
>>> ITE TIM <itetim@socketcity.com> 01/20/00 05:12pm >>>
Date: Thu, 20 Jan 2000 16:08:16 -0600 (CST)
From: Chuck Miller <cmiller@vuse.vanderbilt.edu>
Subject: Temporary freeway detection/surveillance
I an curious whether any agency has used temporary detection and/or
surveillance equipment to monitor traffic in construction zones on
freeways.
</snip>
Top
-------------------- 2 --------------------
#2. TIM: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY -Reply - from
Clifford Conklin
Top
Date: Fri, 21 Jan 2000 21:39:20 -0500
From: Clifford Conklin <cconklin@gw.dot.state.ny.us>
Subject: TIM: ACTION-EMERGENCY VEHICLE LIGHTING POLICY -Reply
My 2 cents. We utilize amber strobes, arrow sticks on our service
patrol vehicles. The strobes work well during the day, but can be
over-powering at night, so it is up to the driver whether he turns on the
strobes at night, most do. Also we prefer that they use the arrow
sticks as much as possible and this seems to work well.
Cliff
Top
-------------------- 3 --------------------
#3. RE: ACTION-EMERGENCY VEHICLE LIGHTING POLICY -
from O'Laughlin, John B.
Top
Date: Sun, 23 Jan 2000 23:13:10 -0500
From: "O'Laughlin, John B." <OLAUGHLIN@pbworld.com>
Subject: RE: ACTION-EMERGENCY VEHICLE LIGHTING POLICY
Phoenix Fire Department has recently put out a training tape that quotes
studies in Penn. Cal. and New York. They advocate reduced lighting
to amber only at night and no strobes or wig wag headlights when at sites.
I'm not sure the titles or dates of the studies.
Top
-------------------- 4 --------------------
#4. Re: TIM: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY -Reply - from Ken Brooke
Top
Date: Mon, 24 Jan 2000 09:14:56 -0500
From: Ken Brooke <ken@mitretek.org>
Subject: Re: TIM: ACTION-EMERGENCY VEHICLE LIGHTING POLICY -Reply
....and this isn't even worth two cents, but thought it might be worth at
least a chuckle.
Our (Fairfax County Fire & Rescue) engines are all fully equipped with
emergency lighting, per NFPA standards, with one addition. The
county fathers decided several years ago to mount amber arrow boards
midships to be used by the engine driver as a primitive means of traffic
control. (We also carry flags and cones, as well as a box of
flares.) Much to the surprise of our county, a state police trooper
promptly ticketed one of our engine drivers for illegal lights. It
seems that state law forbids mounting emergency lights on anything except
duly authorized emergency vehicles, but also forbids mounting amber
direction lighting on emergency vehicles. Apparently that was
reserved for DOT vehicles. It was resolved somehow in a manner
unknown to me, but nowadays the state law remains unchanged, but the
troopers are no longer ticketing us poor put-upon engine drivers.
Our EMS vehicles, and some of the special-purpose vehicles also have
high=power amber directional lighting. I haven't heard them called
arrow sticks. They consist of a line of rectangular amber strobes
with an arrow-shaped one on each end, which can be made to sequence left,
right, or split both ways. They are not as powerful as the white/red
versions, and don't tend to blind you as much. I have no idea how
effective they are, or how clearly they communicate a message to
approaching motorists. I could imagine some confusion between
interpreting them as turn signals or traffic control signals.
-Ken
Top
-------------------- 5 --------------------
#5. TIM: RE: Video: Why Can't We Talk? -Reply -
from Helman, David
Top
Date: 24 Jan 2000 09:59:28 -0500
From: "Helman, David <FHWA>" <David.Helman@fhwa.dot.gov>
Subject: TIM: RE: Video: Why Can't We Talk? -Reply
There are two messages in the video - one direct and one indirect.
The overt message of the video is one of spectrum. The hidden
message is one of interoperability. My first reaction watching the
video was "so when are they going to get to the message I want to hear?"
They never do, but the message of inter-agency coordination is implied.
Top
-------------------- 6 --------------------
#6. Re: TIM: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY -Reply - from Dick Raub
Top
Date: Mon, 24 Jan 2000 09:56:14 -0600
From: Dick Raub <raraub@nwu.edu>
Subject: Re: TIM: ACTION-EMERGENCY VEHICLE LIGHTING POLICY -Reply
Ken:
State laws, can too often be viewed with a delightful laugh coupled with a
HUH????
One problem. No matter how screwy and laughable, they somehow find a
way to bite us.
Dick
Top
-------------------- 7 --------------------
#7. RE: TIM: ACTION-EMERGENCY VEHICLE LIGHTING
POLICY -Reply - from
O'Laughlin, John B.
Top
Date: Mon, 24 Jan 2000 11:53:26 -0500
From: "O'Laughlin, John B." <OLAUGHLIN@pbworld.com>
Subject: RE: TIM: ACTION-EMERGENCY VEHICLE LIGHTING POLICY -Reply
In most states, lights and other emergency equipment are covered by
administrative laws or procedures. They can be changed without
legislative involvement. In any states that have trouble with
reconfiguring lights or other equipment, the best approach is request a
change in the policy before installing the lights. Usually the state
has one organization that takes care of this process and it could be the
state police or general administration. This is another good
opportunity for Incident Management task forces or teams to get something
accomplished.
In our workshops around the country, C D Tyler, Ted Smith and I have been
pushing for dialog on lighting. Phoenix fire, as usual, has taken
the lead on this subject. They are very well respected in the fire
business and when they do something like this, other fire departments will
give it a fair review.
We have also gotten feedback from previous workshops that it works!!!
Rural Arizona officers, fire departments and DOT have combined to try this
and were pleasantly surprised how well traffic smoothed out when they cut
off all emergency lights except the rear amber lights on the rear
emergency vehicle. When they shut off most lights, the motorists
apparently think what ever was happening is over and they don't gawk
nearly as much.
I had one major towing company tell me they wanted to put arrow boards
on their recovery equipment to assist at scenes with traffic management
and were told by a DOT they couldn't because arrow boards were for DOT
only. And here I thought that kind of possessive thinking was reserved for
us old police guys!!!
Top
-------------------- 8 --------------------
#8. TIM: Traffic Incident Management Congestion
Ratios - from Corbin,
John
Top
Date: Mon, 31 Jan 2000 15:26:21 -0600
From: "Corbin, John" <john.corbin@dot.state.wi.us>
Subject: TIM: Traffic Incident Management Congestion Ratios
-----Original Message-----
From: George Saylor [mailto:gsaylor@dot.state.oh.us]
Sent: Wednesday, January 26, 2000 10:25 am
To: DOT@LISTSERV.NODAK.EDU
Subject: Incident Management Summary
List Members:
When I posted my request for information regarding current practices for
quantifying incident-to-recurring congestion ratios I was unaware of the
importance of this subject to others. I received many detailed
responses to my inquiry in addition to several requests to summarize the
information for this group.
In the interest of semantics, I will refer to accident or other
non-recurring delay/congestion as "Incident" and normal day-to-day Level
of Service types of delays as "Recurring".
I. The very first response provided a comprehensive resource
document - the 1999 Annual Mobility Report prepared by TTI. The
report (in PDF) can be found at
http://mobility.tamu.edu/ This report provided me with data for
Cleveland, Cincinnati, and Columbus, OH and in the format I had requested
- Incident vs. Recurring Delay. An interesting observation of the
Ohio data is the ratio of incident-to-recurring delay being almost 10%
lower than other urban areas of a similar population group. However,
the Ohio cities had aggregate recurring congestion almost 50% higher than
the large urban area average. The numbers seem to indicate that
these cities have lower accident rates but much higher volume-to-capacity.
However, I wouldn't rely upon that conclusion exclusively to explain the
numbers.
II. Another response indicated that a major information source of
the Mobility Report is the FHWA HPMS data. The respondent also
suggested that there are many influencing factors but no clear indication
of what significance these factors play in the results. For example,
duration of the peak period and duration of incidents certainly influence
congestion measurement. This response also included an interesting
observation - a police officer in a small urban area mentioned during a
workshop that their incident-to-recurring congestion ratio was 100%,
meaning that unless there was an incident, there was no congestion.
III. California DOT responded that they had conducted a study
several years ago which identified a 50-50 split in incident-to-recurring
congestion. The current situation is one where focused efforts on
incident management, along with overall congestion growth, indicate a
probable increase in recurring congestion.
IV. Cambridge Systematics, Inc. has prepared a report for FHWA -
"Sketch Methods for Estimating Incident-Related Impacts", December 1998.
This report used recurring bottlenecks per mile in assessing the influence
of recurring vs. non-recurring congestion on a daily basis. Table
2.1 in the report illustrates the relationship between volume/capacity
groups, presence of useable shoulders, and the number of bottlenecks per
mile.
V. Vermont DOT performs a high accident analysis for roadway and
intersection programs. Since this state is primarily rural, accident
locations are quantifiable, however, recurring congestion is more a matter
of perception.
VI. Battelle is conducting an ATIS customer satisfaction survey
for FHWA in three of the MDI areas. There appears to be some
correlation between frequency of ATIS use and the amount of unexpected
congestion.
VII. Colorado DOT has developed an incident management measure
for reporting to their Transportation Commissioners. The measure is
the number and duration of road closures by fiscal year. This takes
into account weather and congestion-related causes for the closure.
VIII. The PATH Program has an evaluation known as the Freeway
Service Patrol Project
http://www.path.berkeley.edu/FSP/
There is quite a bit of data associated with this study which provides
some enlightenment.
IX. Although not a specific response to my inquiry, I did receive
information recently through another source which is somewhat relevant.
The Institute of Transportation Studies at the University of California
has web page
http://www.its.berkeley.edu/projects/freewaydata/
which provides a number of papers concerning using loop detectors to
derive travel time and the onset of congestion.
*******************************
My comments:
I. Surprise No. 1 - Not only the number but the variety of
responses. Surprise No. 2 - The level of interest in a summary of the
responses. It would appear that there's no "easy" answer this
time in my efforts to provide my superiors with "THE Solution". Many
agencies (or so it would seem) are struggling with the issue of incident
management and the quantifiable elements of congestion.
II. The distinction between recurring congestion and
incident-related congestion is not always clear-cut. Responses to
the survey indicate that this distinction can be subjective due to the way
we measure congestion. For the sake of argument, I'll categorized
incident data as being of three types - data for public consumption, data
for public sector policy needs, and data for fairly intense technical
analysis. The responses offered examples of all three.
III. Do incident detection algorithms in freeway management systems
really detect incidents? Do we need to rethink the thresholds of
congestion to better identify causes? My inquiry attempted to elicit
information from ITS/freeway management systems (FMS). This was
motivated by an interest in how others deal with the distinction between
incidents and recurring congestion in real-time. My opinion is that
Traffic Management Centers have provided a highly competitive arena for
automated incident detection algorithms. Some algorithms can take up
to 5 minutes to detect an "incident". Newer methods being
investigated have gotten this response time to under a minute. But
is that still good enough? Statistically, what are the chances that
an operator will be visually cued, detect, and verify, an incident faster
than an algorithm?
IV. Are we attempting to do too much with real-time data
collection? Or are we not doing enough? I envision the future
of Traffic Management Centers as being more automated with much less human
intervention. My casual observation has been that this may be
difficult to achieve. Not only are we still struggling with the
analytical issues, but the status quo for real-time operations
necessitates a high degree of operator functions. Also, I would like
to see more, and better access to archived data from real-time systems for
research and analysis purposes. May be it's already there but I just
can't find it.
V. The responses I received have provided me with information I
might have otherwise overlooked. So, in a sense, an "Answer"
was provided. I now have better information with which to use in my
efforts.
Thank you all.
George Saylor
ODOT
----------------------------------------------------------------------------
George W. Saylor, P.E., Transportation Systems Engineer
Ohio DOT - Office of Urban & Corridor Planning
1980 W. Broad St., Columbus, OH 43223-1102
(614) 752-8099 Fax (614) 728-9358
----------------------------------------------------------------------------
Top
-------------------- 9 --------------------
#9. Re: TIM: Traffic Incident Management Congestion Ratios
- from Ken Brooke
Top
Date: Tue, 01 Feb 2000 11:16:29 -0500
From: Ken Brooke <ken@mitretek.org>
Subject: Re: TIM: Traffic Incident Management Congestion Ratios
I found Mr. Saylor's note to be very interesting. I hope he is
subscribed to the list, as I cannot find an email address for him.
Over the years, I've also done analyses using sensor data, but I've never
dealt with traffic information. So please take my comments below
FWIW, and my own surprise as from one used to more formal and consistent
methods. I'm fairly new to traffic systems, and certainly still
learning. Thanks to Mr. Saylor for making me realize that I may have
skipped a whole level of understanding. I've again found a new
dimension to my ignorance.
This connects with my small part of the ITS world. Our work in
the IEEE 1512 (ITS Incident Management Message Sets Standard) Working
Group largely presumes that incident detection/verification has occurred
prior to the commencement of incident management. We and the message
sets could still benefit from using standardized descriptors of an
incident.
Think of it this way. An incident requires some resources to
respond and remedy the situation. In order to request the dispatch
of the appropriate response resources, one must describe the
characteristics of the incident to the dispatcher. Then, the
dispatcher must pass this info along to the responders. I believe
that it would be most effective to incorporate the best descriptors that
qualitatively and quantitatively describe the situation. Such
information may have been ultimately derived from sensor data, and should
be consistently used.
These would presume standards of use (and design, of course). So,
everyone should know what is meant by "heavy traffic" and "xx lanes
blocked" and "time of occurrence", for example....right?
I think there may be opportunity for original work to be done regarding
the semantics of the "what", "when", and "how long" of highway incidents.
Then again, maybe it's all already been adopted, and just not evident in
these few study examples. The related semantics regarding the
question of "where" are probably best handled in a separate thread.
Any LRMS folks tuned in?
Ken Brooke
---------------------------------------------------------------------
ITE TIM (That is George W. Saylor)wrote:
Top
-------------------- 10 --------------------
#10. Re: TIM: RE: Portable Cameras - from
Larry Corcoran
Top
Date: Tue, 01 Feb 2000 12:54:59 -0700
From: "Larry Corcoran" <LCorcora@douglas.co.us>
Subject: Re: TIM: RE: Portable Cameras
The Colorado DOT's Traffic Management Center (CTMC) built an Incident
Command/Response Vehicle a couple of years ago with a roof mount for a
color/tilt/pan/zoom CCTV that fed the picture into the inside office where
there was a monitor and VCR. It also had the capability to send the
picture back to the CTMC over the cell phone. The unit had low band,
VHF, UHF and 800 MHz radios as well as regular cellular phones and
satellite cell phone capability. The unit did send a compressed
video picture back to the CTMC for one fire department mass casualty
training exercise and a real incident where an overhead sign bridge was
knocked partially down by a dump truck with the bed up.
They may not be still using it for incidents but still have the vehicle.
A contact would be Frank Kinder at (303) 512-5820.
Top
-------------------- 11 --------------------
#11. RE: TIM: RE: Portable Cameras - from
Bruce Churchill
Top
Date: Tue, 1 Feb 2000 13:33:55 -0800
From: "Bruce Churchill" <bchurch@la.nateng.com>
Subject: RE: TIM: RE: Portable Cameras
Bill:
You can also check with WA DOT and Indiana DOT. WA uses Incident Response
Trucks that I believe are now testing a remotely operated CCTV unit. POC
would be Jerry Althauser, (206)726-6752 e-mail:
althaug@wsdot.wa.gov. I'm not
sure about the IN DOT contact but I bet John Corbin would know that one.
A similar project was submitted for funding to the Southern CA Priority
Corridor SHOWCASE program but it never made the cut. The project was to
put a remote CCTV on a Motorist Aid Patrol unit operated by Samaritania
Inc. who does a lot of service patrols on the East Coast - the video would
have been transmitted back to the Caltrans District 8 TMC in San
Bernardino from the beat units on a remote stretch of I-15 from Barstow to
the Nevada state line. We figured that a standard cellular modem would be
sufficient to transmit video at a rate of about 1 frame per 2-3 sec. If
CDPD had been available we could have done better than that. A company
that has experience with wireless transmission of CCTV video is Comarco
Wireless Technologies in Irvine, CA. A POC there would be Sebastian
Gutierrez (sgutierrez@comarco.com)
or Kevin Gonzalez (949)788-2830.
Bruce Churchill
NET Corporation
La Mirada, CA
Top
-------------------- 12 --------------------
#12. Re: Traffic Incident Detection - from
Ken Brooke
Top
Date: Wed, 02 Feb 2000 11:50:51 -0500
From: Ken Brooke <ken@mitretek.org>
Subject: Re: Traffic Incident Detection
I hope you won't mind that I forwarded your reply to the ITE Traffic
Incident Management LISTSERV, as that was where I picked up your original
message. John Corbin (copied on this message) could probably provide
you with the ways and means of subscribing to the list. I think your
input would be welcomed in the discussions there.
Questions for the group:
If there actually is a need for a common vernacular relating
incident/non-recurring congestion phenomena to traffic sensor data, how
should that be established? Perhaps we can benefit from the
scientific community's learning experience in this area. Looking
back on my own limited experience with psychological and medical
experimentation data, I see a heavy reliance on statistical analysis tools
for hypothesis testing. Is that appropriate for detecting the
occurrence of incidents by analyzing traffic sensor data?
Ken
Top
-------------------- 13 --------------------
#13. TIM: Re: Traffic Incident Detection - from
Barry Drogin
Top
Date: Wed, 02 Feb 2000 14:02:22 -0500
From: "Barry Drogin" <BDrogin@mtabt.org>
Subject: TIM: Re: Traffic Incident Detection
** Reply Requested When Convenient **
You wrote:
If there actually is a need for a common vernacular relating
incident/non-recurring congestion phenomena to traffic sensor data, how
should that be established?
There is a need for a mathematical middle ground, so that different forms
of sensors and descriptive input can be characterized internal to an
incident management system, that info spread to other systems, and then
each system analyze the data and help implement responses, predict traffic
flows and be updated as the incident "clears". This mathematical
middle ground could take the form of a system state matrix (traffic
volume, traffic flow, highway capacity, time of day, type of day) along
with incident-specific parameters (lane blockage, medical emergency,
removal means).
With all the data flowing in, it is absurd to seek a "common vernacular"
that reduces an incident down to agreed on definitions of "severe",
"major", "minor" and so on (a scale of one to five?). Fuzzy logic
may be applied as an algorithm, but should not be assumed as appropriate
input or output.
We have the ability to count vehicles and measure vehicle speeds to within
a few percentage points. Surely we can come up with a mathematical
way of describing the impact of an incident on those numbers (say, a
time-varying filter characteristic) that fits our knowledge of traffic and
is more appropriate to the new tools we are using? Then we can set
thresholds and come up with precise definitions for vague terms we've been
using for years.
Barry Drogin
ATMS Project Manager
ATMS Program Management Division/Engineering & Construction Dept.
MTA Bridges & Tunnels (Triborough Bridge and Tunnel Authority)
Robert Moses Building/Randall's Island
New York, NY 10035-0035
bdrogin@mtabt.org
(212) 870-6535
(212) 870-6515 fax
(917) 641-0533 pager
Top
-------------------- 14 --------------------
#14. Re: TIM: Re: Traffic Incident Detection -
from Ken Brooke
Top
Date: Wed, 02 Feb 2000 15:35:21 -0500
From: Ken Brooke <ken@mitretek.org>
Subject: Re: TIM: Re: Traffic Incident Detection
You said it better than I. My "common vernacular" is your
"mathematical way of describing the impact". I should hasten to
clarify that I would not propose standardizing universal incident
criteria, only standard ways of describing them. I once fancied
myself a statistician, and it seems to me that statistical analysis still
offers the best, time-proven way of doing what you describe. I think
that by using such a disciplined approach, defending one's conclusions
should be much easier. I don't know any other way of normalizing
sensor data sampled from different traffic flows to enable valid
comparisons or aggregations.
A sensor-measured traffic incident in downtown Lancaster, PA would exhibit
very different characteristics than an incident on the Woodrow Wilson
Bridge in Washington, DC. And yet, it seems such information seems
to get merged without adjustment. Bushels of apples and oranges, to
be sure....or maybe applesauce and marmalade.
-Ken Brooke
Top
-------------------- 15 --------------------
#15. Re: TIM: Re: Traffic Incident Detection -
from Barry Drogin
Top
Date: Wed, 02 Feb 2000 16:00:11 -0500
From: "Barry Drogin" <BDrogin@mtabt.org>
Subject: Re: TIM: Re: Traffic Incident Detection
** Reply Requested When Convenient **
I'm an old information theory man myself, but I see correlation functions
based on real data more as proof than as generative. Rather than
posit a set of inputs and a set of outputs and try to derive statistically
the connection between them, I think it would be better to propose a set
of interim system state variables based on theory and correlate inputs to
the state variables and outputs to the state variables. That way,
with incomplete input data, one can still estimate the system state
variables and generate best estimate output as a result.
Some of these system state variables may be geographic constants. So
if an identical set of inputs from two different locations give different
outputs, it may be that an additional system variable is needed once the
difference is identified (rather than more sensors). For example,
one location may have shoulders but the other doesn't, and this is shown
to have an effect on how emergency vehicles can be located. When an
incident is located, data is fetched from a GIS database and loaded into
the appropriate system variable.
Barry Drogin
ATMS Project Manager
ATMS Program Management Division/Engineering & Construction Dept.
MTA Bridges & Tunnels (Triborough Bridge and Tunnel Authority)
Robert Moses Building/Randall's Island
New York, NY 10035-0035
bdrogin@mtabt.org
(212) 870-6535
(212) 870-6515 fax
(917) 641-0533 pager
Top |