Welcome to J4mZzy’s personal website!
About Me
My first two years of research focused on developing tools for coordination, communication, and safety for multi-agent systems with rich heterogeneity, where agents try to accomplish different tasks. To coordinate such systems, we leveraged techniques from sheaf theory, allowing us to formulate an optimization problem whose solution is the optimal trajectory and input for each agent. To allow agents to solve this problem in an online fashion, we established a family of distributed optimization algorithms that are robust to asynchrony. To ensure agents do not collide while executing the optimal input, we developed a framework for designing and deploying time-varying control barrier functions to guarantee safety.
My future research plans include perception-based motion planning, with a particular interest in graph and affordance-based planning. I aim to develop planners that reason over structured scene representations, such as affordance graphs derived from perception, to enable robots to identify and exploit semantic structure in their environments when computing safe, dynamically feasible trajectories. I am interested in bridging the mathematical foundations from my earlier work, including sheaf-theoretic modeling, barrier function methods, and convex geometric representations, with these perception-driven planning approaches for real-world autonomous navigation and manipulation.
I welcome collaborations. Feel free to email me for academic discussions.