25,248 sensors for La La Land.
Afternoon rush hour traffic is streaming down La Cienega Boulevard until it hits an unexpected bottleneck at the intersection with Olympic Boulevard in Beverly Hills. Two police cars, lights flashing, are parked by a bus stop, blocking a vital lane. Delays are slowly but unmistakably building up at the traffic lights, causing the sensors in each lane to send silent alerts to a nondescript building in downtown Los Angeles.
That’s when Jeffrey Xu springs into action. The veteran transportation engineer is in charge of ATSAC, the Automated Traffic Surveillance and Control System that keeps the city’s 7.5 million vehicles moving. As soon as the line of red dots across all lanes at this major artery pops up on Xu’s screen, he routes the intersection’s camera feed to a wall of monitors in front of his desk and pans around. “Police activity. Let’s adjust the cycle time a bit more,” he says and overrides the software to manually allot a few more seconds to drivers on La Cienega in order to make up for the blocked lane.
The busy intersection northwest of downtown LA is one of almost 4,700 signals that ATSAC’s algorithms constantly monitor and adjust, taking into account variables such as realtime traffic flow, travel speed, time and day of the week, special events and holidays, even fire trucks racing to an emergency. It’s a sophisticated electronic nervous system that tries to watch and optimize the way cars and trucks, pedestrians and bikers, as well as buses and light rail lines get around this Southern Californian metropolis with 7,500 miles (12.070 km) of surface roads, not counting the notoriously clogged Freeways which would add another 181 miles (291 km).
After more than three decades of development and expansion dating back to the first Olympics held in LA in 1984, ATSAC has become the most advanced traffic control system in the U.S., covering 468 square miles (1,212 square km). It digests a growing list of data sources, primarily second-by-second input from more than 25,000 sensors embedded in the asphalt and 560 cameras watching key intersections each reading travels via a hardwired connection to the ATSAC control room four floors below a downtown LA office building across City Hall. The system also keeps tabs on almost 35,000 parking meters to monitor which spaces are taken and can adjust their rates according to demand.
Los Angeles has long hovered near the top of global traffic jam rankings, with close to six million daily commuters taking to the streets. According to the congestion index complied by navigation services provider TomTom, the city ranks tenth in the world and first in the U.S. for its congestion, increasing average traffic time by 41 percent and bumping annual fuel consumption by 25 gallons (95 l) per car. Angelenos spend about 43 additional minutes a day stuck in traffic, which adds up to almost a full week wasted behind the wheel in a year. That means a cross-town trip from Santa Monica to downtown LA covering 16 miles (27 km) takes more than an hour during peak times. Although its notorious highways are not part of the ATSAC network, the 405 Freeway is a particularly dire symbol for LA’s gridlock. More than 300,000 cars crawl down the nation’s busiest highway every day, leading transportation experts to extend the definition of “rush hour” to a full five hours, lasting from 3 to 8 pm.
Bikes are harder to detect than cars
Engineers with the city’s Department of Transportation (LADOT) developed the algorithms in-house that reconcile current conditions with a historic baseline for each location at a given point in time. It enables the system to automatically adjust parameters such as the cycle time for a signal, the offset between multiple intersections so drivers catch a green wave, and giving priority to individual buses that run behind schedule. For more complex intersections it incorporates data from light rail, metro or dedicated bus lanes, as well as from the growing network of bike lanes with their own signals. The latter proved particularly difficult to engineer, since humans on a bike are harder to detect as a car. Since two members of Xu’s team are only staffing the control center from 6 am to 7 am Monday through Friday, the system runs automatically on weekends and during the night.
“We built ATSAC in two major phases, beginning with the original vision dating back to 1984, and continue to fill it in to be ready for the coming world of the Internet of Things, connected vehicles and autonomous driving,” explains Dan Mitchell, Assistant General Manager at LADOT. The last milestone was achieved in 2013 when every traffic light in the city was connected to one of the 40 local hubs that form a redundant network on a fiber optic link feeding into ATSAC’s central brain. To date, ATSAC has cost roughly $420 million in federal, state and local funds, and it has been adopted by other cities.
' We want to see the whole film as vehicles move around. '
Jeffrey Xu, Supervisor at ATSAC
The current system is working as intended, according to researchers at the Texas A&M Transportation Institute. They found that ATSAC improves overall travel speeds by 13 percent, decreases average stop delays by at least one-third and reduces fuel consumption and emissions by 3-4 percent. So even as congestion might creep up due to the growth in population and traffic, ATSAC prevents even worse congestion and pollution.
But the biggest changes lie ahead as traffic keeps growing and the entire region is aiming for a more multimodal mix away from individual cars. Officials have spent billions to build more than 100 miles of passenger rail to better connect LA County, and voters last November approved a measure to invest $42 billion in additional rail and bus rapid transit lines. The long-term hope is to convert a quarter of the county’s population into regular transit riders, more than triple the current rate. Add to that the rise of ridesharing fleets and autonomous vehicles, and LADOT manager Mitchell sees the opportunity to create a data-driven and therefore more intelligent and seamless mobility experience.
Feeding signals directly into a car’s dashboard
He envisions a feedback loop between the data ATSAC gathers and individual users via smartphone apps or connected vehicles that would add actively or passively crowdsourced data to the mix. A partnership with Google-owned mapping service Waze, which an estimated 1.5 million Angelenos use to avoid traffic jams, is a first step in that direction.
A live feed of slowdowns or accidents reported by Waze users now shows up in the ATSAC control center to alert engineers to emerging problems. “Up until now, we’ve only had information for discrete places, like snapshots, but we want to see the whole film as vehicles move around,” explains ATSAC supervisor Xu. As more cars become connected to the city’s infrastructure, his agency could also feed live information from traffic signals directly into a car’s dashboard, for instance to recommend optimum travel speed to drivers.
“You’d be challenged to find another city with so much data that’s accessible from any point. We have a unique opportunity to share that data and optimize traffic flow,” says Mitchell. Among other things, he envisions third-party providers to work with the city to offer new services around a better transportation experience. The Department of Transportation this past fall spelled out its vision for ATSAC 3.0 to evolve into an “innovation platform” for “urban mobility in a digital age” in order to improve pedestrian safety and create smooth multi-modal connectivity.
' You’d be challenged to find another city with so much data that’s accessible from any point. '
Dan Mitchell, Assistant General Manager at LADOT
Visions of efficient traffic
This vision fits with a larger trend to make cities smarter for everyone on the move. Mapping platform HERE, for instance, just announced a new feature called “electronic horizon” that feeds hyperlocal traffic information directly into high-definition maps in a vehicle so driver assistance or future autonomous systems can anticipate what’s immediately ahead. And in Pittsburgh, a startup spun out of Carnegie Mellon University called Surtrac is building its own network of smart traffic lights. Equipped with radar and cameras, they make their own timing decisions and then share those with surrounding signals. The next phase calls for sharing that information with surrounding cars. It’s a different approach than ATSAC’s centralized model, but racing toward the same finish line: making traffic more fluid and efficient.