UNM professor leading the future of safe autonomy
Dr. Meeko Oishi’s work on human-automation systems is helping solve problems in a vast array of disciplines such as robot navigation, aircraft flight management systems and space vehicle docking.
Oishi, an associate professor and Regents’ Lecturer in the Department of Electrical and Computer Engineering at UNM, collaborates with the Air Force Research Laboratory and Sandia National Laboratories to explore how her methods to could solve various aerospace problems.
“My research focuses on safety and autonomy of systems with a human in the loop. We want to design autonomous systems that can provide assurances of safety and efficiency, despite the complexity and uncertainty associated with a human operator,” Oishi said in a recent interview.
“A common example to describe this would be aircraft flight management systems, which often accomplish tasks that are monitored by the pilot.”
Oishi has spearheaded various projects over her years, but some of her most recent work, on stochastic reachability analysis and controller synthesis, has earned her and her students the highly competitive 2017 IEEE/ACM Hybrid Systems: Computation and Control (HSCC) Best Paper Award and finalist status for the 2018 IEEE/ACM HSCC Best Paper Award.
“My favorite part about this research is that we’re solving a really important yet difficult problem. My students and I are developing both theory and computational tools that could have a real impact in complex, engineered systems.”– Dr. Meeko Oishi
Stochastic reachability is the process of analyzing the behavior of controlled systems such as spacecraft to assess the likelihood of violation of desired safety properties. Oishi’s work on stochastic reachability analysis has focused on developing numerical methods that improve the computational cost for certain classes of systems, by exploiting mathematical properties.
Oishi said the generic form of the calculation is based in dynamic programming, which is typically prohibitively expensive. By creating new mathematically derived computational tools, the cost can be dramatically reduced.
“Due to my students’ work here at UNM, we can now solve problems that used to take days or weeks to solve … in just seconds,” — Dr. Meeko Oishi
Oishi said UNM is a very enjoyable and unique place to work and do research.
“I believe that the opportunities for collaboration both within UNM as well as with the national laboratories nearby really makes UNM unique. At UNM you are only limited by what you can dream up.”
Follow her progress and other research topics here.