US cities use AI to help ambulances, fire trucks fight traffic to arrive faster

There’s technology to analyse roads crowded with vehicles, letting a fire truck driver know the best route to take to a fire. PHOTO: REUTERS

WASHINGTON - With bogus 911 calls and clogged streets making it harder for ambulances and fire engines to get where they are needed, overburdened US municipal departments are finding they can get through jams with technology that breaks down traffic patterns and categorises emergencies.

US cities have become a testing ground for start-ups to improve emergency response times – and they promise to save money while doing so.

Research group C2Smarter is using sensors to help fire engines get through Manhattan’s snarled streets faster, while Lyt is deploying artificial intelligence (AI) to decide when to turn red lights green in San Jose and Seattle, so vehicles get to their destinations more quickly.

MD Ally has found a way to triage and redirect 911 calls in Phoenix and Fort Myers, Florida, so hospital emergency rooms are not inundated with non-critical cases, while getting people the care they need.

This follows a general trend of AI-driven productivity gains and corporate cost-cutting across the US but in these cases, the positive results can save lives.

In New York City alone, the average emergency response time has grown to almost eight minutes, more than a minute longer than it was in 2013, while response times for fires have also lengthened.

C2Smarter – a federally funded consortium of researchers from seven institutions headed by New York University’s Tandon School of Engineering – has begun work to create a digital twin of some streets in Manhattan’s Harlem neighbourhood, mimicking traffic patterns to improve the fire department’s response times.

“We know traffic is getting worse, it’s affecting EMS (Emergency Medical Services) and fire response times and we want to understand better why that is,” says New York City Fire Commissioner Laura Kavanagh.

“How can we make recommendations unless we understand what human behaviour is behind it? Not only where is it happening, but also what impacts it and what mitigates it.”

C2Smarter’s technology piggybacks on existing street sensor data to analyse roads that have grown more crowded in recent years with delivery e-bikes, Amazon trucks and curbside restaurant sheds. It can then tell a truck driver the best route to take to a fire.

Lead researcher Joseph Chow hopes to do real-time simulations, testing out, say, which vehicle might be in the best location to respond to an emergency.

Depending on the vagaries of everyday traffic, it might not be the one closest to the scene. “We’re still in the lab stage,” says Dr Chow, whose team is getting all the data in place to train the AI-based traffic prediction system. By the summer, “we should have something more substantial”.

Founded in 2016, Lyt got its start helping city buses move through their routes faster, but its “transit signal priority” technology can be applied to first-responder vehicles as well.

Founder and chief executive officer Tim Menard started out as an engineer at Toyota and then Tesla, working on traffic simulators and smart cars. He says of what he thought at the time: “If we can just bring all that technology back to cities, we can change traffic for everybody.”

He created a system to connect the sensors that track city vehicles with the software that controls traffic signals, feeding it real-time data on street conditions, which Lyt’s AI analyses to adjust the timing of red and green lights, and get buses to their next stop quickly.

The system, which learns patterns as it goes, can even be applied to municipal snow ploughs.

San Jose started testing Lyt on a handful of bus routes in 2018, followed by Seattle, Portland, Oregon and other cities.

In San Jose and Portland, buses have spent 9,107 fewer hours stuck at red lights: that is 54 weeks of hassle saved. Less idling also saves fuel – worth US$569,000 (S$777,000) in 2023 in those two cities alone – and lowers carbon dioxide emissions.

Over the next two years, San Jose plans to expand Lyt to 242 more intersections, while Toronto is using it to optimise its routes for snow ploughs. “We’re jamming pretty well right now,” he says.

Traffic is not the only thing that delays ambulances and fire engines. About half of the 240 million calls made to 911 each year are not emergencies, according to the entrepreneur behind MD Ally, Ms Shanel Fields, whose father was a volunteer emergency medical technician.

That gave her an early look at 911 systems, which in many parts of the country are understaffed and overtaxed by non-critical calls from people who do not understand when to use the system or who rely on it for routine healthcare needs.

“911 is the first number we learn to call as children and many people use it to get the care they need,” says Ms Fields, who earned her MBA at the Wharton School of the University of Pennsylvania in 2019. “So 911 has expanded to care for communities in a wide variety of ways and there is only one response option – the emergency one.”

Unnecessary calls flood hospital ERs with people who do not need to be there. The Massachusetts General Hospital ER has hit critical levels of crowding nearly every day for the past 16 months, creating what its president, Dr David Brown, has called “a full-blown crisis” that jeopardises patient care.

Those unneeded ambulance trips and ER visits cost an estimated US$31.5 billion a year.

MD Ally integrates with 911 dispatch and electronic patient-record systems to shuttle low-urgency calls to a telehealth video session. There, callers speak to an ER doctor for up to an hour, going over medical histories and medications to determine a treatment plan, along with any necessary referrals to specialists.

“We answer within 15 seconds and provide someone to talk to within 15 minutes,” Ms Fields says. Afterwards, a “care concierge” from MD Ally follows up to schedule appointments and can even book an Uber ride to the visit.

Payment for these services varies by state and depends on the patient’s insurance coverage.

MD Ally now serves more than five million people in Arizona, California and Florida, with a waiting list of more than a dozen communities.

Florida’s Lee County, which includes the city of Fort Myers, gets about 350 emergency calls a day. It used to rely on a local hospital network for telehealth assistance, but the service got inundated during the pandemic by people afraid to enter hospitals.

“The nice thing about MD Ally is they have this blanket system that works for any agency,” says Mr Colin Johnson, Lee County’s deputy chief of EMS.

More than the technology, though, it is the human touch that has impressed Mr Johnson the most, like when Ms Fields personally picked up medication for a patient and delivered it to his home, then later worked with state officials to move him to a care facility.

“That individual, sometimes he called five to seven times a day,” Mr Johnson says. “We’ve not had a single call since. That’s where it matters.” BLOOMBERG

Join ST's Telegram channel and get the latest breaking news delivered to you.