Safety-Critical Control for Human-Robot Interaction in Shared Workspaces using Control Barrier Functions

The transition from isolated industrial automation to collaborative environments where humans and robots operate in shared workspaces presents significant challenges in ensuring safety without compromising operational efficiency. This paper investigates the application of Control Barrier Functions as a rigorous mathematical framework for synthesizing safety-critical controllers in human-robot interaction scenarios. Unlike traditional safety protocols that rely on stopping mechanisms or wide separation distances, the proposed approach allows for continuous, seamless operation by formally guaranteeing that the robotic system state remains within a forward-invariant safe set. We formulate the control problem as a Quadratic Program that mediates between a nominal performance-driven controller and safety constraints derived from the barrier functions. This formulation enables the system to deviate from its intended trajectory only minimally and only when necessary to prevent imminent collision. We address the dynamic uncertainty inherent in human motion by incorporating robust barrier formulations that account for worst-case velocity bounds. The theoretical underpinnings are validated through high-fidelity simulations and comparative analyses against potential field methods. The results demonstrate that the control barrier function approach significantly reduces jerk and trajectory oscillations while maintaining strict safety guarantees, thereby facilitating smoother and more productive human-robot collaboration.