Control Design with Guarantees for Minimalist Robotic Swarms
Himani Sinhmar,Postdoctoral Research Associate,Princeton University,
Despite significant advances in multi-agent systems, coordinating teams in unstructured GPS-denied environments with no direct communication remains challenging. This talk addresses key questions: How minimal can an agent's abilities be while guaranteeing desired behavior, safety, and robustness? What quantifiable trade-offs arise between task, system, and control design parameters? I will show how we can design provably correct reactive controllers for a multi-agent target search and encapsulation in a dynamic, unknown environment. The controller relies solely on a robot's simple sensors to construct an impression of its surroundings. I will also provide theoretical guarantees for encapsulating faster evaders where each agent lacks knowledge of the evader's motion, relaxing a strong assumption in pursuit-evasion literature. I will finally present a derivative-free control algorithm for target encapsulation in cluttered environments. This research lays the groundwork to design scalable minimalist robot swarms tailored to specific tasks, ensuring guaranteed behavior and safety across diverse applications such as nanomedicine and aerial robots for wildfire control.
