Kristy Sakano

Hi!👋🏼 My name is Kristy and I’m a third-year Ph.D. student in Computer Science at the University of Maryland, College Park, advised by Dr. Mumu Xu of the Department of Aerospace Engineering. I’m also part of the GAMMA Lab in the Department of Computer Science which is where you’ll find me most days.

I’m broadly interested in formal verification and model checking for autonomous cyber-physical systems that operate in messy, real-world environments. My work combines human feedback, Signal Temporal Logic (STL), and vision-language models (VLMs) to encode informal human rules as formal specifications and enforce them at runtime 🤖

I care about building systems that are verifiable, explainable, and grounded in human intent, and my current research revolves around:

Integrating human feedback into formal verification pipelines 💬
Using VLMs to bridge natural language and safety logic in autonomous robot navigation 📸
Runtime enforcement of temporal logic specifications ⏱️
Evaluating robot compliance with real-world, human-defined safety standards 📝

This past summer (2026), I interned at the Department of Transportation Highly Automated Systems Safety Center of Excellence (DOT HASS COE), where I developed a pruning and matching algorithm for historical autonomous vehicle (AV) crash data from the NHTSA SGO and the California DMV. We conducted exploratory data analysis to identify trends that could support automated crash typology classification and integrate the two datasets, providing a more comprehensive understanding of the circumstances leading up to AV crashes. This work ultimately aims to identify gaps in our understanding of AV safety and inform efforts to prevent future crashes.

Before my PhD, I worked at NAWCAD Patuxent River (2018-2024) as an Operations Research Analyst and Autonomous Systems Scientist, where I collaborated as a visiting scientist with NASA Goddard (2021) and interned at the Naval Research Laboratory (2019). I earned my MS in Computer Science with a focus on Computational Perception & Robotics at Georgia Tech through their Online Master of Science in Computer Science (OMSCS) program in 2023 🐝, and a BS in Physics with Astrophysics concentration from the University of North Carolina at Chapel Hill in 2018 🐏

One fun fact: during my undergrad, I conducted research at the Harvard-Smithsonian Center for Astrophysics through a Smithsonian-funded program, studying pulsating star systems. That work remains my most cited paper to date 🌟

news

Aug 27, 2026	This past summer, I wrapped up my internship at the USDOT HASS COE and presented our work titled From Language to Logic: A Theoretical Architecture for VLM-Grounded Safe Navigation at ICUAS 2026 in Corfu, Greece! A copy of our paper is available in the conference proceedings here (Note that the site is password protected, UN and PW both ‘ICUAS2026’.) 🎓 We also published our recent work, DRIVE-SAFE: Data-Driven Robustness and Informed Validation for Evolving Specifications via Formal Evaluation, in Wiley’s Advanced Robotics Research journal. This paper is the second is the second in a two-part series exploring the use of formal methods and temporal specifications to evaluate black-box autonomous mobile robots.📄
May 11, 2026	It’s been a busy spring - three quick announcements below! We presented our work at the High Confidence Software & Systems (HCSS) Conference 2026 this month! 👩🏻‍🏫 Our paper From Language to Logic: A Theoretical Architecture for VLM-Grounded Safe Navigation was accepted to ICUAS 2026 - come see our talk at the Testing and Evaluation: Autonomy II track on Thursday, June 18th! 🛩️ Our research made the UMD Aerospace Engineering Spring 2026 newsletter! Read more about it here.🛣️
Feb 20, 2026	I am excited to join the US epartment of Transportation Highly Automated Systems Safety Center of Excellence (DOT HASS COE) this coming summer as a DOTSI intern!
Oct 16, 2025	The Xu Research Group participated in the AI for Good Media Day at UMD College Park!
Oct 09, 2025	MarylandToday wrote an article about our research! Check it out here.

selected publications

ICUAS25

A Framework for Black-Box Controller Design to Automatically Satisfy Specifications Using Signal Temporal Logic

Kristy Sakano, Joe Mockler, Alexis Chen, and 1 more author

In 2025 International Conference on Unmanned Aircraft Systems (ICUAS), 2025

Abs DOI Slides

We present a framework for including humanreadable specifications in the design of black-box autonomous systems, or systems whose construct are prohibitively complex to analyze or intuit. By integrating parametric Signal Temporal Logic (pSTL), we can systematically evaluate and refine deep reinforcement learning policies to ensure compliance with predefined system specifications. Our approach is tested in a simulated autonomous driving environment, where we train a deep reinforcement learning agent in Mario Kart SNES using Proximal Policy Optimization. The agent is evaluated based on its ability to navigate a structured driving course while satisfying a set of a priori requirements; this evaluation is performed by writing and solving the parameters in a pSTL statement. This work contributes to the broader effort of bridging formal methods and data-driven learning, providing insights for researchers and operators alike in developing AI based controllers for real-world autonomous systems.