Cal Poly alumnus Christopher Young vividly remembers his first search and rescue mission on Nov. 5, 1984. Missing person Roberta “Bibi” Lee mysteriously disappeared, prompting Young and the Contra Costa Search and Rescue Unit to search Redwood Regional Park in Oakland Hills, CA.
The search was not successful, and Lee’s body was found a month later in the park, according to Young. However, Young’s experience in this case led him on a path to becoming an industry expert in search and rescue operations with over 40 years of experience.
“It’s a classic mystery like Sherlock Holmes,” Young said. “I can’t say specifically what it was that evening that got me hooked, but it was the fact that it was just, ‘Hey I’m out here looking for somebody,’ and that’s kind of the motivation that got me started.”
Young now lends his expertise to Cal Poly’s AI for Search and Rescue, an ongoing three-year student project that aims to apply technology in search and rescue missions nationwide. He helps the team understand operation logistics in a search and how to leverage artificial intelligence to dig deep into clues to find missing persons.
“The more we get into this [project], it’s infectious, everybody wants to make this happen,” Young said.
Artificial intelligence is not a new concept, but its recent resurgence has taken the technology industry by storm, according to WIRED.
Young said that time equals life or death. The project aims to replace mountains of paperwork with technology and artificial intelligence to increase efficiency. By analyzing historical data of past search and rescue missions, artificial intelligence finds patterns where people are usually found and predicts where the missing person might be. The patterns inform heat maps that hone in on a precise location.
Privately funded by Cal Poly alum Gary Bloom, the project took off in summer 2021 and has fostered student research, senior projects and master’s theses at Cal Poly, according to a presentation for a conference for missing persons in Las Vegas in April.
“If we can compress the time that it takes to find the missing person [with technology], then we’re saving lives and making a difference in the outcome of the individual and their family, friends [and] coworkers,” Bloom said.
Bloom first started working for search and rescue units when he was a student at Cal Poly and has seen first-hand the consequences of finding missing persons too late. He said that starting a small pilot program using technology to gather data on a search and rescue mission made him realize the possibilities of predicting where a missing individual can be found.
The project consists of five teams: frontend development, backend development, deep learning, machine learning, and probabilistic reasoning teams, with the last three implementing artificial intelligence.
Ideally, during a search and rescue mission, searchers will log information about the missing person in real time into the project’s database, allowing its technology to determine the location of the missing person in as little time as possible, according to Siddarth Viswanathan, Cal Poly alumni, and team leader for probabilistic reasoning.
With a passion for AI technology, Viswanathan said he was drawn to the project’s potential for good and began working on it his freshman year.
“This project is one that you can clearly see how it’s being used and how it’s benefiting others,” Viswanathan said. “That’s always something that I really like, because I’m not just working on a project that’s helping me gain knowledge.”
As for fellow project member Julian Duran, the computer science senior took a “leap of faith” and joined the project since he felt it has a positive impact on people.
“When somebody gets lost, there’s this instinct that people want to go find them,” Duran said. “What’s cool about search and rescue is that you don’t have to be trained; untrained people can be a part of it and they can actually help.”
Duran leads the frontend development team, focusing on creating a user-friendly website that search and rescue organizations can use to organize data. The goal is to streamline data collection and access it easily, he said.
Profit is not a driving force for the team, according to Young. Since the majority of search and rescue organizations are volunteers, the project team plans to make their project available to search and rescue teams nationwide.
“We may not be the ones to see it come to fruition, but we have certainly been the catalyst that made it happen,” Young said.
Since AI for Search and Rescue is a student-led project, Bloom said that is the reason why they may not see the project come to fruition. Students graduate, and other younger students step in to continue their work. The “continual turnover” of new students and the “labor of love” is what keeps them working on the project, Bloom said.
“Even if we never deliver the system, we will have proven what’s possible: [that] AI can be used to solve this problem,” Bloom said. “We’re really pursuing the art of what’s possible.”
The project will be tested in mock searches throughout the summer, with the team ready to conduct field tests for feedback and improve their technology to find and save people.
Now 40 years after his first search mission, Young often wonders if having the technology that the students are creating now would have found Lee faster. He said that he wishes that his team could capture data and analyze it faster.
