Los Angeles

The city of Los Angeles is famous for its car-centric culture and notorious for the epic bumper-to-bumper traffic jams resulting from it. The city’s sprawling nature makes the build-out of public transportation networks complex, with busses often stuck in the same traffic as single-occupancy vehicles.

Now, however, the Los Angeles County Metropolitan Transportation Authority, commonly referred to as Metro, is tapping into the power of artificial intelligence (AI) to help clear a path forward for riding one of the most extensive networks in the United States.

As first reported in the Los Angeles Times, Metro is installing AI-enabled cameras onto the fronts of busses, which will take photos of the license plates of vehicles illegally parked in bus lanes.

The $11 million program, incorporating 100 camera systems, is scheduled to go live at the end of this year after a test phase this summer.

To address privacy concerns, non-violation footage is required to be deleted within 15 days and violation footage retained for up to six months, or 60 days after citation disposition.

The project is the result of a deal Metro stuck with AI technology company Hayden AI’s “Automated Bus Lane Enforcement” platform, which utilizes cameras equipped with AI algorithms to monitor designated bus lanes in real-time.

The AI algorithms analyze the video footage captured by the cameras to detect violations of bus lane rules, such as unauthorized vehicles driving or parking in the bus lane.

Authorized agencies gain access to potential violations through Hayden AI’s data portal, which integrates with citation processing systems. When a violation is detected, the system automatically generates citations or alerts for enforcement agencies.

The platform can also provide valuable data insights and analytics to transportation agencies, including information on the frequency and location of violations, which can be used to optimize bus lane design, enforcement strategies and traffic management.

Pedro Pacheco, a Gartner analyst focused on the mobility industry, said that while the technology is not new—multiple AI-aided cameras can be found on the latest passenger vehicles—use of the technology in service of public transport could have a range of positive benefits.

“There could be countless applications of what you can do with computer vision, because a camera is basically picking images of a particular situation,” he explained. “You could also train these cameras to detect how many people are waiting for the bus, to detect crime, or even the cleanliness of the bus stop.”

The implementation of automated bus lane enforcement, as observed in New York City since 2019, has shown significant increases in speeds on enforced bus routes, contributing to the reduction of greenhouse gas emissions, traffic congestion and improvement of public health.

Following Washington, DC’s lead, Los Angeles also plans to deploy Hayden AI’s automated bus stop enforcement to curb illegal parking, with the aim of ensuring safety and accessibility for all riders at bus stops, particularly those with disabilities.

New York City’s MTA automated bus lane enforcement program has notably boosted compliance with dedicated bus lane restrictions, with 86% of violators not repeating offenses, signaling a shift in driver behavior.

Moreover, the program has led to enhanced safety on deployed routes, evidenced by a 34% decrease in collisions along the M15-SBS route since its implementation.

Pacheco said he agrees the use of AI-enabled cameras can improve safety by offering operating organizations more data and visual information about traffic incidents.

“They could figure out what caused an emergency stop, for example, and then use that data to improve driver training,” he explained.

By networking all these computer vision cameras together, route operators could also improve fleet management by assessing traffic patterns, and improve punctuality or identify bottlenecks.

Taking the potential capabilities one step further could lead to vehicle to vehicle (V2V) communication between the busses and other cars on the road, using data the bus has accumulated to help drivers find parking spots nearby instead of clogging up bus lanes.

“One approach is obviously to punish the wrongdoers, but a more proactive way would be to leverage all the real time information the bus is gathering and share that with other vehicles to improve the overall flow of traffic in the area,” Pacheco said.