February 3, 2004
Project Members and Core Competencies
Sean Savage: Writing, Design and Prototyping.
Lauren Wilkinson: Writing, Software Design and Database
Programming.
Mikhail Avrekh: Information Architecture and Database
Programming.
John Han: Technical Policy, Writing, Software Design
and Programming.
Road sensors placed along highways throughout the Bay Area provide reams of up-to-the-minute data and a warehouse of historical traffic data that’s in the public domain, but plenty of this potentially priceless information is going to waste because it’s inaccessible to regular people. There are a couple of Web sites that provide traffic visualizations based on this data, but the data on these sites is presented in an extremely low level of granularity; a user can glance at these images to get a vague idea of how bad traffic is throughout the Bay Area, but nothing is available that tailors this information to a user’s individual routes, and nothing actively suggests alternative routes based on this data. On the other hand, mapping service such as Mapquest provide personalized direction maps at a high level of granularity, but these resources don’t take into account traffic data. Depending on time of day, traffic can render a Mapquest-recommended route one of the worst possible routes between two points. We propose a system that combines real-time traffic data, traffic forecasts based on historical data, map visualizations and routing services to provide users with a new service that suggests what truly are the best routes and that also allows users to visually explore alternative routes and alternative travel times.
We expect core users to include:
dispatchers who will use the service for assistance in
coordinating taxi drivers, delivery drivers, police, firefighters and
ambulances;
recreational drivers who use the system to explore and
choose routes to follow during occasional road trips; and
daily commuters who will use the system to minimize aggravation
and time spent getting to and from work.
We expect to find participants through:
friends and acquaintances who (a)
drive for weekend trips or (b) commute daily using Bay Area freeways;
and
referrals provided to us through
transportation contacts we have already made in the City Planning &
Civil Engineering Departments at Berkeley, and the Transportation Center
at University of Washington.
Initial Design
The initial design for the information architecture is as follows:
The backend consists of an established group Oracle database, local to SIMS, containing archived traffic data obtained from the Civil Engineering department. This Oracle database receives real-time data updates regularly. We have chosen to port the data to a local Oracle account to maintain independence from the Civil Engineering project goals.
The middleware includes a complete, functional statistics
package. This package includes:
traffic-based routing intelligence; and
traffic prediction intelligence, based on aggregate patterns
in archived
data plus real-time data.
We are streamlining middleware that has already been developed by the
Civil Engineering folks.
The front-end is a Web-based application that includes
the following functionality:
an interactive map that displays
data from the Oracle database;
optimal routing based on real-time
and historical traffic data; and
optimal departure time, given a
fixed route and desired arrival time, based on traffic data.
Initial Scenarios
Sample initial scenarios we envision:
DISPATCHER/DRIVER SCENARIOS
The user is a dispatcher who guides delivery people in the Bay Area. The
delivery people drive throughout the workday in the Bay Area, along inconsistent
routes. The driver calls in to dispatcher to learn her next destination,
the dispatcher responds with that destination as well as with the best
route to follow in getting there. The dispatcher uses a standard PC with
a full-scale Web browser and a high-speed Internet connection to use the
service. The system quickly shows the dispatcher what’s the best
route for the current driver based on current traffic conditions, and
also based on historical data that might suggest upcoming changes in the
conditions.
RECREATIONAL DRIVER
A driver in Berkeley wants to go to Lake Tahoe tomorrow. He has never
made this drive and he isn’t familiar with the route. He wants to
decide when to leave in the morning, in order to arrive by a certain time.
After learning the user’s starting and destination points, and his
preferred departure time, the application suggests a preferred route and
departure time, and it also shows a few alternatives.
DAILY COMMUTER
A professional who commutes between San Francisco and San Jose each day
can log in at any time to view the current state of her routes. She can
log in just before leaving work, or just before leaving home, to see a
route whose traffic visualizations indicate current, up-to-the-minute
traffic data along that route, modified slightly by historical data projections.
If the system expects an alternate route to be more efficient today, it
will highlight this route and warn the user to avoid the usual route.