My research focuses on decision-making and control for robots operating in dynamic, interactive environments, especially under limited sensing, computation, and communication. My goal is to build robots that can think and act simultaneously, using physical action to elicit useful information, resolve ambiguity, and continuously adapt their decisions as they interact with the world. I study how to design the underlying decision and control mechanisms so that these interactions give rise to individual and collective behaviors that we can shape and rigorously certify. My work draws on control theory, nonlinear dynamical systems, computational neuroscience, and collective behavior, with the broader goal of building autonomous systems whose behavior is safe, adaptive, and interpretable by design. I am currently working with Prof. Naomi Leonard at Princeton. I did my PhD at Cornell University with Prof. Hadas Kress-Gazit, and my Bachelor's and Master's in Aerospace Engineering at IIT Bombay. Outside of research, I enjoy running, hiking, and reading about world affairs, psychology, and philosophy of science.
I study how robots can make fast, reliable, and provably safe decisions when sensing, computation, communication, or actuation are limited. My work combines control theory, nonlinear dynamics, and collective intelligence to develop decentralized decision-making and control frameworks that are both mathematically grounded and deployable on real robotic systems.
Decentralized, game-theoretic, and neuromorphic control for scalable environment monitoring in resource-constrained robot teams.
Safe, scalable, and deadlock-free multi-robot navigation through continuous adaptation and local interaction rules.
Provably safe aerial motion planning under uncertainty, limited computation, and complex workspace constraints.
Control, estimation, and modeling for distributed spacecraft systems, autonomous navigation, and propulsion.
Provably correct decentralized control for robot swarms with no memory, no communication, and no localization.
Micrometer-scale origami robots that fold into 3D shapes, locomote in solution, and are controlled by surface electrochemical actuators.
Reactive task and motion planning using object affordances, feasibility checks, and tool substitution — demonstrated on a Stretch robot.
Advised students are marked with † and equal contributions with *.