(Bloomberg) -- The Central Station monorail platform hangs above the streets of Jacksonville, Florida. The fareboxes, which haven’t accepted payment in almost six years, wear vinyl wrap with “SKYWAY IS FREE” stenciled in white block letters. Even at that price, the end of a recent morning rush finds only a handful of riders bothering to wait for the silver-and-blue trains that glide in every six minutes.

On a cool day in January, the wait might not be much of an impediment. But it’s too long during sweltering Florida summers, when air-conditioned monorail cars become an oasis, says Brad Thoburn, vice president of planning, development and innovation at the Jacksonville Transportation Authority. It’s part of the reason he believes the monorail, a failed relic of a future that never arrived, needs to be replaced by the next sweeping vision of urban transit: autonomous vehicles.

The Skyway was supposed to mature into a sleek system moving millions of riders each year from Jacksonville’s core to a historic district on the north side, what’s now a stadium for the city’s National Football League team to the east and beyond. But the monorail never left downtown after its 1989 debut or extended past a two-and-a-half mile track. Daily ridership hovers around 5,000, just 10 percent of initial projections.

Pulling down the elevated tracks or mothballing the system would trigger payback of tens of millions of dollars to the federal and state agencies that bankrolled most of the construction costs. Which is another reason why the JTA has pinned its hopes on driving robots.

The authority’s new five-year plan calls for off-ramps to be built at key points along the monorail route, allowing the self-driving shuttles down to reach the street in dedicated traffic lanes. That will make it more affordable for the stunted transit system to finally expand without erecting new tracks—and it would make Jacksonville one of the first cities to tie its transit fate so closely to automated vehicles.

In case the hopeful dynamic isn’t clear enough, Thoburn’s office is decorated with a version of the “Field of Dreams” movie poster that casts him as the main character, who hears a voice intoning, “If you build it, they will come.”

Jacksonville’s initiative, dubbed Ultimate Urban Circulator, is the most visible endorsement yet of autonomous vehicles by a public agency at a time when cities around the world are just beginning to experiment with the nascent technology. “We’re not throwing our hands up,” said Nathaniel Ford, JTA's chief executive officer. “We researched transportation infrastructure and technology around the world, and yes—no surprise—autonomous vehicles are going to be the future.”

Converting a rail system in this way has no real precedent, and its success will hinge to a large extent on forces outside JTA’s control. Yet to planners in Jacksonville, the benefits that autonomous vehicles will bring are clear.

A larger fleet of individual robo-shuttles can provide service that adjusts according to demand patterns. Vehicles can arrive at stations more often than today’s six-minute gaps, or a convoy could be dispatched to handle peak demand at a single location. And by utilizing both the elevated track and the streets below, the system can finally reach vibrant neighborhoods outside Jacksonville’s downtown.

Perhaps best of all, JTA anticipates that the robot shuttle system’s operating and maintenance costs will be less than the Skyway’s current $6.5 million annual bill.

But Jacksonville’s plan will work only after cities master automated driving systems. More than 1,000 pilot projects have been launched worldwide to evaluate how new technology might fit into existing transportation networks, and early results have been promising, said Brian Collie, head of the Boston Consulting Group’s automotive practice in North America.

“It’s not surprising why they’re trying to pull this off,” he said of Jacksonville’s effort. “The challenge is the technology is super-close, but it’s not yet there.”

In Las Vegas, an early test site for robotic shuttles, people can catch a driverless ride along a half-mile loop in the city’s historic downtown, thanks to a pilot program sponsored by AAA, formerly the American Automobile Association. The shuttle is operated by a French transportation service company, Keolis SA, in a yearlong evaluation.

Being one of the first to try it means that Las Vegas gets the distinction of being one of the initial locations in which an unproven technology can collide with unpredictable humans. The shuttle, made by France’s Navya SAS, didn’t make it through Day One before it was crunched by a tractor-trailer backing into an alley.

Another pilot program at the University of Michigan in Ann Arbor is currently testing two Navya driverless shuttles. In this case, the experiment—being run by the university’s MCity vehicle-research lab—is meant to gauge both consumer acceptance and the responses of other drivers. After a public rollout slated for this spring, the shuttles will ferry students, faculty and staff along a mile-long route on public roads between an engineering and research hub on campus.

Tech and automotive giants, including General Motors Co. and Alphabet Inc.’s Waymo self-driving unit, are racing to launch urban robotaxi services, which loom as powerful would-be competitors to Jacksonville’s future shuttles. Waymo has racked up more than 4 million self-driving miles to date, and the company is gearing up to offer a commercial automated ride-hailing service in Phoenix later this year.

Jacksonville, meanwhile, is still in the early days of testing. The robotic shuttles that might one day replace the monorail are put through their paces on a short test-and-learn strip in a parking lot near EverBank Field, home of the Jaguars. On a recent evening an EasyMile EZ10 vehicle embarked on the straight course. The shuttle crept quietly between stops, guided by “virtual rails” coded into its computer.

“The basic idea is that we’re putting multiple types of autonomous technologies to the test and, at the same time, getting the public on board,” said Matthew Chang, JTA’s program administrator for the project. “We’re going to make a very large purchase of a fleet of vehicles, so this is our chance to get it right.”

The EZ10 is equipped with lidar sensors to map the world around it and detect unexpected objects. At one point, a staffer from Transdev SA, the French transportation-systems company working with JTA, stepped into the shuttle’s path, causing it to come to a stop.

The plan calls for two years of further testing, but already the shuttles have shown clear limitations. The EZ10 and other vehicles under consideration aren’t fully compliant with U.S. disability access laws, which JTA must follow. The current vehicles aren’t durable enough to be used in public transportation. And the agency wants vehicles that can fit more than the 12-person capacity of the EZ10.

“It’s not a question of if they will perfect it—it’s just a when, in terms of perfecting these technologies,” JTA’s CEO Ford said.

The Jacksonville plan has other major uncertainties, too, including how it will be paid for and whether automated vehicles will live up to expectations. That’s not to mention whether enough people, who are still mostly afraid of riding in self-driving vehicles, will actually use it.

Use of the monorail is at such a low ebb that there’s almost nowhere for ridership in Jacksonville to go but up. In 2010, the Florida Times-Union published a column deriding the monorail system in a headline: “After 20 years, Jacksonville Skyway Remains a Punchline.” The joke was not lost on Ford when he took over two years later, after running more robust transit authorities in San Francisco and Atlanta.

“When I arrived here, it definitely was” a punchline, he said. “It didn't really get to where people wanted to go.”

A public review of modernization options presented several non-starters. Extending the elevated track would cost up to $100 million per mile. Replacing the aging trains would cost at least $4 million each. Ending the service or tearing down the tracks would trigger repayment obligations that JTA officials estimate to be roughly $50 million.

The entire annual operations budget for the JTA is about $200 million. Betting on the untested robots seemed one of the only realistic options.

“We’re taking aging, 30-year-old infrastructure that our taxpayers invested in—the federal government invested in,” Ford said, “and we’re marrying it with new technology.”

When Jacksonville’s monorail was conceived in the 1970s, it was the new technology. Planning coincided with a small wave of similar “peoplemovers,” a once-futuristic concept backed by a deep-pocketed federal program to fund urban transportation. Detroit and Miami were among the other cities that received grants from the Urban Mass Transit Administration, which is now the Federal Transit Administration. An early plan from 1976 envisioned 49,000 office workers, tourists and residents gliding around downtown Jacksonville each day on graceful trains poised above snarled traffic.

Now that future is worn out. Just six of the system’s 10 monorail trains remain in service. The other four are being cannibalized for parts, since the manufacturer, Canada’s Bombardier Inc., no longer sells the same model. Ford thinks the trains have about five years left.

But the elevated track, with a 50-year lifespan and a steep repayment penalty for disuse, must remain active through at least 2039. “If you look at the cost of the aerial structure and to go to the neighborhoods where we’re going to, it literally would be a billion-dollar project to expand what we currently have,” Ford said.

Autonomous shuttles appear to be a relative bargain. The vehicles cost around $250,000 to $300,000, according to Ford, with lower prices expected as the technology matures and economies of scale kick in. That’s about as much as the JTA chief will reveal about costs. Work on a more complete estimate is underway but, he cautioned, the unprecedented nature of the plan makes even rough projections difficult.

Who will pay for it? Ford has proposed a mix of federal funds and private-sector partnerships, including development deals around new stations or the sharing of future fare revenues.

“I would be curious as to what that looks like,” said Ginger Goodin, director of the Transportation Policy Research Center at Texas A&M University. “I’m not seeing a lot of private-sector companies step forward.”

Jacksonville’s approach raises further practical issues. Creating dedicated lanes for autonomous vehicles would squeeze traffic into less space, Goodin said, risking a backlash from drivers. “What they’ve done in gathering public input, understanding travel needs and trying to marry all of that with a technological solution, that’s a great start,” Goodin said. “You’ve got to have that kind of a vision to be able to move to the next step.”

The challenges aren’t lost on Ford, and in some ways he remains inspired by the boldness of his monorail-planning predecessors. “Think about the risk they took 30 years ago, when they first conceived the Skyway,” he said. “I think there’s a lot less risk in what we’re doing today.”

Bloomberg News