Event
Future Leaders in Robotics and AI Seminar: Himani Sinhmar
Friday, April 25, 2025
2:00 p.m.-2:30 p.m.
https://go.umd.edu/FutureLeaders
Yancy Diaz-Mercado
301 405 6506
yancy@umd.edu
Provably Scalable and Decentralized Control Design for Minimalist Robot Swarms
Himani Sinhmar
Postdoctoral Fellow
Princeton University
Abstract
How much collective intelligence can emerge from individually simple agents? In this talk, I explore this question through the lens of minimalist robotic swarms—systems composed of robots operating without communication, GPS, or global reference frames, using only minimal sensing and lightweight onboard computation. I present a control-theoretic framework inspired by opinion dynamics—originally studied in social systems— adapted here for real-time robotic decision-making. By evolving internal dynamical states based solely on local observations, agents achieve emergent behaviors such as consensus for coordination and dissensus for resolving conflicts, including deadlock, in dynamic, partially observable environments. A core contribution of this work is the mathematical characterization of performance guarantees and design trade-offs: for example, how many agents are needed to offset limited sensing? How do sensing constraints impact coordination speed and safety? What minimal assumptions ensure correctness? These questions define a principled design space linking theoretical guarantees to practical considerations such as cost, robustness, and scalability. Beyond navigation, I extend this framework to distributed resource allocation, showing how multi-option opinion dynamics enable decentralized task selection under uncertainty, paving the way toward robust, learning-free swarm algorithms. If you're interested in how minimalist principles can yield scalable, reliable swarm behaviors without relying on centralized control or intensive computation, this talk will offer both the theoretical foundations and practical tools to get you there.
Biography
Himani Sinhmar is a Postdoctoral Research Associate in Leonard's Lab at Princeton University. She received her PhD from Cornell University (2024) and her bachelor's from IIT Bombay (2019). Her research focus is to develop mathematical models of collective intelligence, revealing fundamental principles of how natural and artificial groups (such as animal flocks and robot swarms) coordinate complex behaviors. Her work bridges theoretical guarantees of collective performance with practical, hardware-compatible robotic designs. Sinhmar’s research contributions have been recognized through awards such as the Robotics: Science and Systems Pioneers and Caltech's Trailblazing Young Researcher Award.
