Graph Neural Network-Based Safe Multi-Robot Control

Multi-robot systems must satisfy strict safety requirements such as collision avoidance and minimum mutual distance. Control barrier functions (CBFs) are the standard tool for guaranteeing these constraints in real time, but designing a good CBF is difficult and often leads to overly conservative — and so lower-performing — behaviour.

This thesis explores a learning-based approach to improve CBF performance for multi-robot collision avoidance, potentially factoring in sensor noise and communication delays, with simulations in ROS and Gazebo.