How does Fort Worth check how many of its streetlights are burned out? It sends city staff into the dark — literally.
Staffers pair up quarterly to drive around city streets at night and tally the number of lights that are on or off at a given time. That data is used to calculate the “burn rate,” or the percentage of streetlights functioning at a time.
The process is time-consuming and expensive. The city has to pay staffers overtime, and the tallies are only a single-day snapshot. Now, Fort Worth is set to try out a new solution using artificial intelligence.
A portion of the 2024 proposed budget includes $250,000 to create an year-long artificial intelligence pilot program in the Transportation and Public Works Department. The program would take crowdsourced camera footage from vehicles driving around the city, and then analyze it to calculate things like the burn rate without needing to send out a staff member.
The technology has the potential to simplify a variety of tasks, Elizabeth Young, business process manager for TPW, said. The department will set three tests for the pilot program. Calculating the burn rate is one, and assessing pavement condition is another. The third hasn’t been decided yet, but Young said there’s plenty of work that could fit the bill.
“TPW has almost a million assets,” she said. “And so in trying to track all that, we want to know if a sign has been added, if a sign has been knocked down, and so those are potential projects for the future.”
A successful pilot program would help the city become more proactive, rather than reactive, in maintaining its assets, Young said. It would also free up manpower for other tasks, like maintenance projects.
Fort Worth has conducted two large scale inventories of its transportation assets in recent years — one in 2016, and another in 2021. Those inventories focused on pavement conditions, streetlights, poles and pavement markings. By using artificial intelligence assessments, the city hopes to stay on top of inventories without needing to expend a large amount of funds at one time.
“Ultimately, if we can [use AI] for something like pavement condition, instead of trying to hit it every five years, and we can get sort of a snapshot and understand if there is a glaring issue that we might not know about, that would be helpful,” she said. “A lot changes in five years.”
Young said she knows some people may be wary of the city using artificial intelligence. She, too, isn’t all in on the technology as a whole, but believes it’s a useful tool for answering simple questions.
“We’re not changing their lives by simply answering the question, ‘Is the street light on or not?’” she said. “Instead, what we’re doing is freeing up resources to actually go change street lights for them.”
Fort Worth isn’t the first city to experiment with artificial intelligence assessments. Detroit began using technology from a company called RoadBotics to assess its road network in 2019; according to the company’s website, more than 200 governments have used its technology for their own assessments.