portal - Portland State University

gasownerΔιαχείριση Δεδομένων

31 Ιαν 2013 (πριν από 4 χρόνια και 5 μήνες)

211 εμφανίσεις

Paper 271

1

DERIVING METROPOLITAN PLANNING PERFORMANCE MEASURES
FROM A REAL
-
TIME REGIONAL TRANSPORTATION DATA ARCHIVE


Robert L.

B
ERTINI

Associate
Professor

Department of Civil & Environmental Engineering

Portland State University

P.O. Box 751

Portland, Oregon 97207
-
0751

USA

Tel:
+1
503

725

4249

Fax:
+1
503 725
5950

E
-
mail:
bertini@pdx.edu


Steven HANSEN
, Casey NOLAN
, Peter BOSA

Graduate Research Assistants

Intelligent Transportations Systems Laboratory

Portland State Universi
ty

P.O. Box 751

Portland, Oregon 97207
-
0751 USA

Tel:
+1
503

725

4249

Fax:
+1 503 725
5950

E
-
mail:
stevenh@pdx.edu

Email: cnn@pdx.edu

Email:
pbosa@hotmail.com





Abstract:

PORTAL is the Portland
, Oregon

regio
n

s Archived Data User
Service, which arc
hiv
es

freeway sensor data at a 20
-
second level of
aggregation from 4
85

discrete points
. This study uses PORTAL to evaluate
transportation goals set forth by Portland’s Metropolitan Planning
Organization.
This paper first discusses the PORTAL data archive struct
ure,
data storage, data processing and user interface. Next is a discussion of the
specific performance measures developed to assess the performance of the
regional transportation system
: t
he percent of regional highway corridors
exceeding
specified

level

of service standards
is analyzed over space and
time;
the change in average travel times in key corridors is
measured; and
point
-
to
-
point peak and off
-
peak travel times on selected freight significant
highways are analyzed
. This paper details the steps r
equired to accurately
extract these performance measures from PORTAL. The

paper concludes
with lessons learned and next steps for improving transportation performance
measurements using archived loop detector data.


Keywords:

data archive, performance mea
surement, transportation planning,
regional planning, decision support.

Paper 271

2

DERIVING METROPOLITAN PLANNING PERFORMANCE MEASURES
FROM A REAL
-
TIME REGIONAL TRANSPORTATION DATA ARCHIVE

1

INTRODUCTION

The objective of developing a d
ata
a
rchiv
e

for intelligent trans
portation
systems (ITS)
is to s
ystematic
ally

ret
ain

and re
-
use operational ITS data.
It
may be the case that

the original and primary purpose of
produc
ing these
data is for real
-
time
transportation system
management,
the
archiving

of

these
otherwise
-
discar
ded data
can
offer a rich source of information
toward the

evaluat
ion of

system performance
on a continuing basis

(
Turner
2001
). Once
retained,
I
TS data can be used
by

key

stakeholders including metropolitan
planning organizations (MPOs), state transportat
ion planners, traffic
management operators, transit operators and researchers (
USDOT 2000
).

ITS data
can readily

be used to supplement or replace conventional data
sources that are both time consuming and costly to implement, and the high
level of spatial
and temporal detail found in ITS
-
generated data
can
lend itself
to
innovative

methods of analysis (
USDOT 2000
).
The c
ost effectiveness of
existing data collection
programs

is maximized and
the
sampling bias
can be

reduced
due to

the continuous collection
of data, and
can result in
a better
understanding of

the

variability in system performance (
USDOT 2000
).
For
example, transportation agencies invest
i
n

costly

manual traffic count
programs
for data

collect
ion

at a
small

number of sites
for

72 hours
. In
co
ntrast, ITS data are available
from

hundreds or thousands of sensor

locations on a
continuous basis
.
A major

challenge is to
carefully
document
and retain the ITS data so that future users can understand its limitations and
capabilities.

Use of archived d
ata can also assist in diagnosing defective ITS
components and aid in
deployment of maintenance resources
.

To
encourage

the
archiving

of ITS data, the United States Department of
Transportation (USDOT)
has developed
the Archived Data User Service
(ADUS)

as

part of
the National ITS Architecture to include “the unambiguous
interchange and reuse of data and information throughout all functional
areas.”
Under the
ADUS
guidelines,
data from ITS systems
are

collected and
archived for historical, secondary and non

real
-
time uses, and
are

made
readily available to users
(
USDOT 1998
)
.


In cooperation with the Oregon Department of Transportation (ODOT) and
other regional partners, the Portland Regional Transportation Archive Listing
(PORTAL, located at portal.its.pdx
.edu) was recently inaugurated via a direct
fiber optic connection between ODOT and Portland State University (PSU). In
July 2004, the data archive “went live,” receiving

a real
-
time stream of

20
-
second data from the 4
85

inductive loop detectors comprisin
g the Portland
area’s Advanced Traffic Management System (ATMS).
This paper first
discusses the PORTAL data archive structure, data storage, data processing
and user interface. Next is a discussion of the specific performance measures
developed to assess
the performance of the regional transportation system:
the percent of regional highway corridors exceeding specified level of service
standards is analyzed over space and time; the change in average travel
times in key corridors is measured; and point
-
to
-
p
oint peak and off
-
peak travel
Paper 271

3



ODOT
TMOC



Loop Controllers
Fiber
PSU
ADUS
Server
Fiber
Web
Server
TriMet
WSDOT
City of
Portland
Future

Backup
Server
Operations
Planning
Public
User Classes

Figure
1
: Portland Metropolitan
Area ADUS Data Flows

times on selected freight significant highways are analyzed
. This paper
details the steps required to accurately extract these

and other sample

performance measures from PORTAL. The paper concludes with lessons
learned and n
ext steps for improving transportation performance

measurements using archived loop detector data.

2

DATA STORAGE AND PROCESSING FOR THE PORTLAND
REGIONAL TRANSPORTATION ARCHIEVE LISTING (PORTAL)

The ODOT Region 1 transportation management operation
s

center

(TMOC)
pulls

data
from

4
85

inductive loop detectors which comprise the Portland
region
's ATMS.
These detectors have been implemented as part of the
region’s

ramp metering system
. Mainline freeway loop detectors are located
upstream of
each
on
-
ramp locatio
n, and the on
-
ramps themselves are also
instrumented.
The

data

generated

(count, occupancy and speed)

are used in
the day to day management of the transportation system to identify congested
areas and incidents and to dispatch incident response or emergen
cy vehicles
to the appropriate locations.
As ODOT upgrades their current pre
-
timed ramp
metering system to one that is dynamic in nature, the quality of these data is
increasingly important for real
-
time operations.

At 20 second intervals, each loop detec
tor records
vehicle count,

the average
speed of these vehicles, and occupancy, or percentage of the sample period
when a vehicle was over the detector. ODOT currently archives 15
-
min
aggregate data for operations control purposes. Prior to implementation o
f
PORTAL
, the raw 20
-
second data were discarded. Now, permitted users can
access and query the archived data through a web interface.
PORTAL’s

overall structure
i
s shown in

Figure
1
. The data are sent from ODOT
TMOC
to PSU in XML format. A script on the PSU server parses the XML data as it
is sent every 20 seconds. The data are stored and accessed using a
PostgreSQL relational database management system (RDBMS) running on a
Linux platform.

The RDBMS stores dat
a physically on a
redundant array of independent disks
(RAID) providing both high
-
speed
access and increased reliability through
redundancy in the event of hardware
failure. The web
-
based user interface is
provided by the Apache web server, with
database c
onnectivity and additional
processing by the PHP hypertext
preprocessor module. Detector locations
have been geo
-
coded for interoperability
with geographic information systems
(GIS).

This archive implements a data
warehousing strategy in that it retains
la
rge amounts of raw operational data
for analysis and decision making
processes, and in that these data are
stored independently of their operational
Paper 271

4

sources, allowing the execution of time
-
consuming queries with no impact on
critical operations uses.

D
ata

security is
a key com
ponent of
the data

archive
, so

t
he working copy of
the database maintained on the primary server is replicated in a compressed
format at a remot
e site. These daily backups
ensure that
PORTAL

can be
rapidly returned to operation with n
o significant loss of data.
T
he primary
database server and backup storage are located in machine rooms with
uninterruptible power supplies (UPS) and generator backup power, preventing
data loss or gaps in data availability due to power outages. The workin
g copy
of the database is stored on a RAID device, providing error detection,
redundancy, and the ability to rebuild missing data upon device failures.
Finally, hardware maintenance and security updates are provided for all
computer systems by experienced
systems administration

personnel
.

2.1

The National ITS Architecture

PORTAL is designed in adherence with

the recommendations of the National
ITS Architecture
, with its

five distinct processing functions (
USDOT 1998
):

1.

Store data in the same format as received f
rom ITS subsystems

2.

Accommodate levels of aggregation and reduction of the data flows,
depending on the type of data represented

3.

Sample raw data flows for permanent storage in accordance with user
specifications. Permanent storage should be either online,
offline, or both.

4.

Apply quality control procedures to the data, including the f
lagging of
suspect data and
editing of data.

5.

Distinguish between the following data types: unprocessed (raw), edited,
aggregated, and transformed.

Raw Data Storage

As loop d
etector data
are

transferred from ODOT TMOC to the PSU relational
database, it is archived with no loss of resolution.
Count
, speed, and
occupancy data are archived indefinitely at 20 second intervals.

Data Aggregation

PORTAL

accommodates levels of aggr
egation appropriate to the type of data
being recorded. Each night at 3 a.m., the 20 second data are aggregated to 5
minute data. The data are also aggregated to 1 hour. Future plans call for
other levels of aggregation

as needed
.

In addition to total vo
lume, average occupancy and average speed, the 5
minute and 1 hour tables also include calculated values of vehicle miles
traveled (VMT), vehicle hours traveled (VHT), delay, and travel time. These
calculations depend on the highway segment length associat
ed with each
loop detector, defined as the segment of the freeway extending halfway to the
next upstream station and the segment extending halfway to the next
downstream station.

Sampling Data

Users

can

sample raw data flows and store data permanently in
accordance
Paper 271

5


Fig
ure
2
: PORTAL Homepage

to their specifications, including aggregation of the data at desired resolutions.
These data can be output on the screen, plotted on a relevant graph, or
downloaded in comma separated value (CSV) format for offline storage.

Quality Control

As
with all ITS data, we experience data quality issues. Fortunately, PSU has
been the first organization to carefully examine and validate the ODOT 20
second loop detector data.
PSUs’ researchers have discovered several errors
through their microscopic anal
ysis of the data. These errors were reported to
ODOT at an early stage of the project.
As a result, ODOT has been
implementing improvements to their system to remove several systematic
errors that were discovered.

Quality control procedures are applied to
identify
and mark suspect or erroneous data.

Distinguish Between Data Types

The final data processing requirement outlined in the

National ITS
Architecture ADUS addendum is that distinctions be maintained between raw,
edited, aggregated and transformed d
ata
. The ADUS user
can query the data
in several

distin
ct areas of the web interface, including time series, data
fidelity,
raw data,
weather, travel time and performance
.

These areas are
discussed further in the next section. In each case the user know
s whether
the query being performed is using raw or aggregated data.

3

DATA RETRIEVAL

Fig
ure
2

shows the current PORTAL user interface. The map shows an

example of how ODOT currently uses loop detector data in real

time. The
highways are color coded to indicate the current traffic speed on each road
segment. The map, maintained by ODOT, is updated every 20 seconds when
the loop detectors record average speed for the time period. This map is also
available on ODOT
’s website www.tripcheck.com along with detailed
information on road conditions.


As shown in the
figure, t
he welcome
screen gives the user
the following options
for querying archived
data, as indicated by
the menu tabs near
the top of the screen:

Time
seri
es
,
Data
Fidelity, Raw

Data,
Weather,

Oblique
Plots, Travel Time,
WIM Data and
Performance.
Selected options are
explained below.

Paper 271

6

3.1

Time
s
eries

The time series tab allows the user to perform the queries on the 5 minute
aggregated data to extract volume, spe
ed, occupancy, vehicle miles traveled,
vehicle hours traveled, travel time, and delay

over a
specific

freeway segment
or for a particular

detector station. Additionally, the user can choose the time
period and specific travel lanes. The user has the opti
on of either plotting the
data, viewing the numbers in a table on the screen, or exporting the data as a
comma separated value file.
Figure
3

shows a
speed contour plot for I
-
5
North on March 16, 2005

and
Figure
4

shows a sample

volume

plot

for one
station along I
-
5 North for March 24, 2005.

Figure
3
:
Sample Speed Contour Plot for I
-
5 North, March 16, 2005















Figure
4
:
Sample Volume Plot for I
-
5
N.

Going
St.

Loop Detector Station, March 24, 2005


Paper 271

7

3.2

Data Fidelity

This section allows the user to determine w
hether the loops queried were
determined to be
functioning properly

errors are flagged due to detector
problems, power or communications outages, planned outages due to
construction and other issues
. The user can either get information on the
health of a
specific detector, or, can get the percentage of healthy loops on a
specified highway section.
Figure 5

shows
an example of the detector health
information available
, which includes a time series contour plot for
Northbound I
-
5 on March 16, 2005, indicatin
g several problem areas and time
periods. Also a pie chart shows the total percentage of “good” readings for
this highway on this day.

This has been
very
useful for assisting ODOT with
the upgrade of some of their traffic management system software and in

updating some of their ramp meter controller hardware. We plan to continue
to refine this component by soliciting input from TMOC staff.





















3.3

Raw Data

The user currently has the option of downloading raw 20 second data in CSV
format for
specific
detector stations or highway segments.

Figure 5
:
D
ata Fidelity I
-
5 North, March 16, 2005

Paper 271

8

3.4

Weather

When conducting traffic studies in the future, it will be important to know what
weather conditions were like on the days analyzed. For example, previous
studies in Oregon have found substantial diff
erences in the numbers of
crashes on rainy days as compared with dry days (
Bertini, R.L
.,
et al

2004
). In
order to provide at least a minimal amount of weather data to future analysts,
PORTAL collects and stores METAR weather observations from National
Oce
anic and Atmospheric Administration (NOAA). Weather observations are
collected from the Portland International Airport
and other sites
. A script
written in Python collects and processes the weather data. The Python
module pyMETAR provides easy access to
NOAA’s weather reports
(
pyMETAR 2004
). Users can query the database to get the weather
conditions for a specified time period. There is a table in the database which
indicates which weather station is geographically closest to each loop
detector. Users
have the option of extracting data from all three weather
stations, or the station that is nearest to a particular loop detector.

Figure 6

shows the hourly temperature and precipitation data for April 1, 2005.
















Figure 6
: Weather Data for Apr
il 1, 2005

4

PERFORMANCE MEASURES

The PORTAL project described in this paper includes the implementation of
a
robust

data archive back end with a state of the art database implementation.
Currently the project has been focusing on the establishment of the fr
eeway
data archive. Additionally, a user
-
oriented front end has been established via
an easy to use web interface. Currently PORTAL has 83 registered users,
mainly representing Oregon’s transportation agencies.
Based on a review of
Paper 271

9

the literature, numer
ous transportation system performance measures have
been developed and tested in recent years for both planning and operational
purposes. Therefore, b
eyond the data archive development, the PORTAL
team is designing, testing and implementing a number of per
formance
measures, with the intent to get them in the hands of transportation system
managers, planners, engineers and even decision
-
makers. Some of these
performance measure implementations are consistent with national efforts to
generate urban congestion

reports and others have been created specifically
for the Portland region. The PORTAL team continually presents sample
measures to regional stakeholders for feedback.


Specific urban performance measures
have

developed to assess the
performance of the reg
ional transportation system

consistent with standards
outlined in the
2000 Regional Transportation Plan (RTP). Examples are given
using
several

key performance measures.
First, daily and monthly
performance reports are described using tabular and graphical

formats.
Second
, the system is examined from the standpoint of measured travel
speed, and differences between time periods are examined graphically. Next,
the concept of peak hour travel time reliability is explored
using

a

map

interface
.

Next
, point
-
to
-
point peak and off
-
peak travel times on I
-
5, a selected
freight significant highway,
are
analyzed.

Lastly, the travel time index and
buffer index measures are described graphically.

4.1

Daily
Reports

Using the Performance tab, users can create daily tabular a
nd graphical
reports for a specific detector station, a highway segment or the entire
freeway network. The report can be generated for specific hours and at either
5 minute or one hour resolution. Measures reported include average flow,
mean travel time,
95th percentile travel time, VMT, VHT, the percentage of
time that congestion is present (defined as 1.3 times the free flow travel
time

FFTT
), the buffer index (95th percentile travel time divided by the free
flow travel time) and the travel time index (a
verage travel time divided by the
free flow travel time).

4.2

Monthly Reports

Similarly, monthly reports are also produced and

Table
1

shows a monthly
report for the entire freeway system for April 2005. As shown, the average
travel

time, travel speed over the segments, measured vehicular speed (along
with free flow values and 95th percentile values) are shown for the entire 144
instrumented miles of freeway. The total VMT and VHT are also tabulated
(and available graphically) and th
e
weather information is further summarized.


As a further example of automated graphical performance measurement,
Figure 7

reveals the average hourly flow throughout the day at 5
-
minute
intervals and also indicates the probability that a 5
-
minute interval

was
considered to be congested during that month.


Paper 271

10

Table
1
: April 2005 Monthly Report

MONTHLY FIGURES

Average
TT

Average
TS

Average
VS

FFTT

FFTS

95th%
TT

95th%
TS


VMT




VHT



Percent

Congested

Hwy
Length

Days
Used


185.
80
Mins



46.53
MPH



54.62
MPH



164.38
Mins



52.60
MPH



232.02
Mins



37.26
MPH



197514974



4147950



13.93%



144.09
Miles


30



WEATHER

NUMBER OF DAYS

Rain ≥ 0.5"

Rain ≥ 1.0"

Fr
o
z
e
n
Precip
itation


Fog

0

0

1

2
















Figur
e 7
: Monthly Report April 2005

4.3

Mapping
Instantaneous Speed

One important measure of transportation system performance is vehicle
speed.
Figure 8

shows an instantaneous speed map, showing speed ranges
for 4:55:00 p.m. on April 18, 2005. This can provide a
traveller

or a manager
with a quick snapshot of freeway conditions at any time.

Paper 271

11















Figure 8
:
Peak Period
Instant Speed
,
April 18, 2005

4.4

Average Speed

With a growing data archive, it is possible to track changes in freeway
performance over time.

As an illustration of this capability,
Figure 9

shows the
average speed measured during peak hours in November 2004, while
Figure
10

shows a similar map for March 2005. In order to examine any changes in
average speed from November to March,
Figure 11

sh
ows a subtraction
between the two previous figures in order to indicate differences in average
speeds from one month to the other. Similar subtraction maps can be used to
assess changes in any of the measured or calculated performance measures.

4.5

Travel Tim
e Reliability

One of the more interesting capabilities developed as part of PORTAL is the
ability to examine travel time trends and reliability. Travelers and shippers
often cite travel time reliability as a major concern, and transportation
agencies are a
ttempting to provide more reliable travel, and better information
about the system reliability in real time.

Following Federal Highway Administration (FHWA) guidelines
, congestion
is
assumed to
occur if the travel time is greater than or equal to 1.3*FFTT
,

where

FFTT is defined as the 15th percentile
t
ravel
t
ime.
Figure 12

shows a
map indicating a measure of travel time reliability for the entire freeway
system during peak hours
. This measure is the ratio of 95th percentile travel
time to mean travel time
for each segment. The measure is indicated by color
code (actually this is the buffer index divided by the travel time index).






Paper 271

12






































4.6

Peak and Off
-
Peak Travel Time Reliability on Freight Corridor

The movement of fre
ight throughout the Portland metropolitan region is taking
on more and more importance, especially with the large increases in truck
flows projected for the next 20 years.
PORTAL

allows a shipper to examine
historical information concerning travel time re
liability on major freight routes
on the network. The northbound Interstate 5 (I
-
5) corridor from Wilsonville to
the Columbia has been chosen as a major freight corridor for this analysis.
Using

northbound

I
-
5 as a

case study
,
Figures 13 and 14

compare

t
ravel time
reliability for an off
-
peak hour with that for a peak hour, during weekdays in
the month of March 2005. As shown, a shipper would find much greater
reliability for this route during the off
-
peak. Information such as this could
prove to be usef
ul to shippers in the Portland region who need to deliver
product within a particular time window.

Figure 9
: Average Speed November 2004


Figure 10
: Average Speed March 2005

Figure 11
: Average Speed Difference

Figure 12
:
Peak Hour Travel Time

Reliability


Paper 271

13























4.7

Travel Time
Index and Buffer Index


The Travel Time Index is the ratio of free flow speed (55 mph) to the average
speed. This
is a ve
ry common congestion
measure

and

is useful for

examining the time periods and locations where speeds fall into congested
conditions
. Figure 15 shows the Travel Time Index for the regional highways.
The Buffer Index measurement is simply the ratio of the fr
ee flow speed to the
95 percentile speed. This measurement is useful for

showing travel time
reliability

for example, this index is an indicator of the travel time allowance
for which a user will be late on average one day per month for a work trip
.
Figure

16 shows the Buffer Index ratings during the month of March for the
regional highways.















Figure 13
:
Point to Point Off
-
Peak Travel
Time Reliabil
ity (I
-
5 N)

Figure 14
:

Point to Point Peak
-
Hour
Travel Time Reliability (I
-
5 N)

Figure 15: Travel Time Index

Figure
16
:
Buffer

Index

Paper 271

14

5

NEXT STEPS

Research across industries indicates that effective data archiving systems
start as small prototypes with a single data source. Studies co
nsistently
indicate that
freeway
detector data is the most useful archived data (
Turner
2001
), and for this reason we have chosen to first archive these data. The
Portland region has a wealth of additional data sources that will be archived at
the appropr
iate time

as part of this research
.
In addition to t
wo data sources

discussed below, PSU is developing an archive of the Portland Regional Land
Information System (RLIS), which contains several land use, land coverage,
and transportation data in a GIS for
mat. The RLIS archive will enable
detection of major changes in land use in the Portland region, and overlay it
with changes in the traffic flow that occurred along freeways, which can be
easily obtained from PORTAL. RLIS was launched in 1995, and is upda
ted
quarterly through Metro
, the Portland regional planning organization
. The
implications of integrating transportation and land use data will provide a rich
source of data, contributing to research investigating how transportation
decisions have been sha
ping land use over time, and vice versa.

5.1

Computer Aided Dispatch Data

Portland’s ATMS includes a comprehensive incident management system
(Bertini, R.L.,
et al
.

2004), which in turn generates a large computer aided
dispatch (CAD) database which then conta
ins information about all recorded
incidents on Portland’s freeway system. This information includes the type of
incident, which lanes were blocked as a result of the incident, and the start
and end time of the incident. The TMOC can record the X
-
Y coord
inates of
an incident, but it is only required if the incident is severe enough to cause
significant delay. In 2001, coordinates were recorded for 10% of all incidents
(Bertini, R.L.,
et al
.

2004). We plan to archive the CAD data so that it can be
further

used to assess the performance of the incident management program
and identify times and locations when incidents affected traffic flow.

5.2

Tri
-
Met Bus Data

TriMet is the transit provider for the Portland, Oregon metropolitan region. As
part of their overal
l operations management system, TriMet has developed a
“smart” bus that enables their automated bus dispatch system (BDS). The
BDS includes Automatic Vehicle Location (AVL) using a satellite
-
based global
positioning system (GPS) on all buses as well as Aut
omatic Passenger
Counters (APCs) installed on a majority of the existing fleet and all new bus
acquisitions. As part of the real
-
time component of the BDS, each vehicle
transmits its location to the transit management center (along with several
other data
points) periodically (roughly every 1
-
3 minutes). These real
-
time
data are currently archived by TriMet.

Each vehicle in the fleet includes an on
-
board memory card that is pre
-
programmed with that bus’ schedule information. Therefore the bus “knows”
where

it is located at all times, and links that with the schedule, enabling a
schedule status (late/early) report to the operator. Further, as part of the
archived data component of the BDS, a data record is written to the memory
card at each bus stop. These

records contain the route number, direction, trip
Paper 271

15

number, date, vehicle number, operator ID, bus stop ID, stop arrival time, stop
departure time, boardings, alightings, and passenger load. Also recorded is
whether the doors of the bus were opened, whethe
r the lift was used, the
dwell time, maximum speed between the previous and current stop, and GPS
coordinates. The memory card data are downloaded at the end of each day
and compiled into a huge database. This information is used by TriMet to
measure syst
em performance on an ongoing basis and to aid in the bus route
scheduling effort that occurs each quarter.

One of the major benefits of TriMet’s real time and archived BDS data is that it
can be used to extract performance characteristics from the region
’s arterial
system. We plan to begin testing the use of both data sources as part of the
Portland archive in the near future for this purpose.

6

CONCLUSION

Although there can be many beneficiaries, data archiving has been historically
and continues to be th
e domain of a single user group (e.g., researchers and
planners) (Turner 2001). The mostly informal arrangements for data archiving
with researchers and planners has occurred due to the inaccessibility and
difficulty in using the large data sets. With the
development of new data
management and processing tools, and the expansion of web
-
based services,
this is now changing. In the Portland, Oregon, metropolitan area, regional
agencies across two states, including cities, counties, MPOs, transit
operators, a
port, a university and state DOTs cooperate by means of
TransPort, the regional ITS coordinating organization. The region’s ITS
Architecture has been developed under the auspices of TransPort.

It can be difficult to determine who should bear the responsib
ility and cost of
maintaining an ITS data archive because the organizations that produce the
data are frequently different from those who benefit from its retention. In
cooperation with the Oregon Department of Transportation (ODOT), TriMet
and the City of

Portland, Portland State University (PSU) is a major user of
regional ITS data for transportation research. Other recent uses of ITS data
by A
DUS stakeholders such as Metro
have been encouraged and facilitated
by PSU researchers already familiar with man
ipulating this type of data.

As a result, PSU is in an ideal position to host transportation data, and its
Intelligent Transportation Systems Laboratory (www.its.pdx.edu) has been
designated as the regional archiving site for ITS data from Portland and
adj
acent areas of southwestern Washington. The design and implementation
of this archive are being carried out in accordance with the functional
requirements for ADUS set forth in the National ITS Architecture, while also
taking into account experiences and s
uggestions from other systems that are
being developed in larger, more populous states and regions (Turner 2001,
USDOT 2000, USDOT 1998, PeMS System 2004)).

This paper has
described the development and implementation of

PORTAL
(see portal.its.pdx.edu)
, a
data archive and performance measurement tool for
the Portland region. In the future we will continue

to evaluate specific
performance measures related to transportation policies implemented by
transportation agencies and decision makers. By actively meas
uring the
Paper 271

16

performance of the system, and tracking this over time, it is hoped that the
Portland region can continue to develop in a way that provides opportunities,
choices and connections in a fiscally responsible and sustainable way.

ACKNOWLEDGEMENTS

The

authors gratefully acknowledge ODOT, TriMet and the City of Portland for
their ongoing support and guidance in the development of the Portland ADUS.
In particular Dennis Mitchell, Galen McGill and Jack Marchant of ODOT; Ken
Turner and Steve Callas of TriM
et; and Bill Kloos and Willie Rotich of the City
of Portland have been instrumental in making this project possible. Dick
Walker, Kyle Hauger and Dr. Gerry Uba of Metro have also enthusiastically
encouraged our efforts. This work is supported by the Nati
onal Science
Foundation, the Portland State University Department of Civil and
Environmental Engineering and the Oregon Engineering and Technology
Industry Council (ETIC). The Portland metropolitan area’s TransPort ITS
committee has also encouraged develo
pment of this project. Andy
Delcambre
, Andy Rodriguez and Spicer Matthews are integral parts of this
project and continue to contribute to its success through their efforts
.

REFERENCES

1.

Turner, S.
(2001)

Guidelines for Developing ITS Data Archiving Systems
.

Report
No.
2127
-
3. FHWA, U.S. Department of Transportation, Texas
Department of Transportation, and

Texas Transportation Institute
.

2.

USDOT (2000)
ITS Data Archiving: Five
-
Year Program Description
.

ADUS
Program
. Washington, DC.

3.

USDOT (1998)
Archived Dat
a User Service (ADUS): An Addendum to the
ITS Program Plan
.
ADUS Program
.

Washington, DC.

4.

Bertini, R.L., Rose, M. and El
-
Geneidy
, A.

(2004)

Using Archived Data to
Measure Operational Benefits of ITS Investments: Region 1 Incident
Response Program
.
Portla
nd State University, Center for
Transportation Studies, Research Report
, June 2004.

5.

pyMETAR (20
04)
http://www.schwarzvogel.de/software
-
pymetar.shtml
.
A
ccessed July 30, 2004.

6.

PeMS System (20
04)
California Department of Transportation and
University of California at Berkeley
.
http://pems.eecs.berkeley.edu/Public/
. Accessed July 30, 2004.