acm-header
Sign In

Communications of the ACM

ACM News

Computer-Automated Traffic Control


View as: Print Mobile App Share:
The Los Angeles ATCAC system control center.

Operators of Los Angeles' Automated Traffic Surveillance and Control system evaluate the flow of traffic on the city's streets, and can alter the timing of signals to improve traffic flow.

Credit: Los Angeles Department of Transportation Communications Office

This past February, the City of Los Angeles added the last pieces of the city's grid to its Automated Traffic Surveillance and Control (ATSAC) system, bringing the total number of intersections governed by the system to 4,398.

ATSAC is one of the largest automated traffic control systems in the world, but it's just one example -- albeit an early, groundbreaking one -- of using computer systems to manage traffic. Other projects in locations from Virginia to Northern California to Korea are helping people find buses, avoid traffic congestion, and locate parking places quicker. These projects are the vanguard of a future in which roads, cars, and traffic signals all communicate with each other to speed traffic and reduce accidents.

ATSAC at first

Los Angeles' ATSAC system began in 1984 with a project to handle traffic around the Los Angeles Coliseum during that year's Olympic Games. As the city obtained additional funding over the years, recounts Jonathan Hui of the Los Angeles Department of Transportation, it added more and more areas, until $150 million from 2008's Proposition 1B provided enough to complete the system.

Each intersection in the ATSAC system is autonomous, Hui explains, with inductive loop detectors embedded in the roadway and a traffic signal controller in a box on the corner. "Traffic signal controllers are pretty old technology that predates the ATSAC system," says Hui. The controllers are wired to area hubs and the hubs to a central control center, where the incoming information is analyzed. The control center originally ran on a mainframe computer supporting workstations, but these days it is hosted on servers with client software running on PCs.

Based on their analysis of the data and images from closed-circuit cameras installed throughout the city, operators at the control center can evaluate the flow of traffic and alter signal timing to improve it. "In many cities, if you want to make signal timing changes, you build a model and test it in the field," says Hui. "But in a city the size of Los Angeles, that's not feasible."

The ATSAC system may be complete, but the work is not done. "There's always congestions, new residential developments, and new patterns in behavior," says Hui. "We're constantly trying to improve signal timing or re-prioritize traffic to accommodate bikes, pedestrians, and so on."

A similar Intelligent Transportation System, likewise built to handle traffic during an international sporting event, is the Parallel Traffic Management System (PtMS) put in place for the 16th Asian Games held in 2010 in Guangzhou (Canton), China. The PtMS has a  video analysis component to detect traffic and passenger flow at stations and on roads, and bus and taxi management modules to allocate public transportation as and where needed. It also has a module that provides information about traffic, events, and weather to travelers, enabling them to plan their time and movements.

Real-time to future time

The ATSAC system and PtMS are prime examples of using real-time data to make operational improvements. "Today, there is lots and lots of data available," says Naveen Lamba, Intelligent Transportation leader at IBM Global Business Services. Traditional sources like loop sensors and newer sources like mobile phones and GPS in buses and taxis can provide "a complete picture of real-time traffic."

"But real time is too late," Lamba continues. When operators have to wait for the data to come in before they can do anything, "you're reacting to what just happened, and it's hard to recover and get ahead of the curve." For that reason, IBM is developing predictive analytics capabilities in its Intelligent Operations for Transportation suite of tools. By combining real-time data with historical knowledge about traffic patterns, says Lamba, "We are able to accurately predict traffic for the next 60 minutes." Based on those predictions, the software can run quick simulations of the outcomes of different interventions, and then present options to the operator.

IBM's system has been deployed to manage traffic flow along the I-15 corridor in San Diego. The Virginia Department of Transportation is launching a project to manage traffic on 35 miles of I-66 leading out of Washington, D.C., the most congested stretch of roadway in the country, with an active traffic management system from TransCore, the company that designed New York City's "Midtown in Motion" next-generation traffic control system.

Making connections

Still, these projects all involve an operator who processes data at a central location and then acts on it. What if the traffic systems could communicate directly with what Lamba calls "connected vehicles"?

One project exploring the possibilities is a partnership between Cisco and "smart parking solution" provider Streetline. Cisco has installed a WiFi network in San Carlos, California (south of San Francisco) that picks up data from Streetline sensors embedded in the roadway. The sensors can tell when a parking space doesn't have a car in it, and communicates that data to Streetline's cloud center. Streetline then pushes the data out to its Parker app for iPhones and Android smartphones.

Cisco's Marc Musgrove points to even more possibilities "as you move from a one-directional network to a multidirectional network," in which vehicles can communicate with the infrastructure and with each other. "Imagine if you could feed information to the cars around you that you have a flat tire, or that you could learn that a car was coming down a side street and wouldn't be able to stop in time," he says. Similarly, a vehicle could communicate its approach to a red light which, if no other vehicle were approaching, could automatically change to green.

Such capabilities will require vehicles equipped with the ability to broadcast such information. "All the auto manufacturers are working towards this vision," says Musgrove. For example, there's a move in the European Union to require all new cars sold after 2015 to have a cellular wireless module for alerting emergency services. "Once you start putting in the network, these are the kinds of things you’ll be able to do," Musgrove says.

Logan Kugler is a freelance technology writer based in Silicon Valley. He has written for over 60 major publications.


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account