Safety-Critical Reactive Motion Using Constrained Variable Admittance Control With Dual-Type Proximity Sensors

We present a method that enhances the safety and responsiveness of robotic manipulators by employing constrained Variable Admittance Control (VAC) in conjunction with proximity perception. Recent studies have shown that manipulators equipped with proximity sensors can effectively avoid nearby obstacles in real-time. Nevertheless, unavoidable collisions remain a critical challenge in human-robot interaction (HRI). As a safety fallback, conventional reactive motion algorithms focus on obstacle avoidance but often suffer from inefficiency and disregard collision handling. Our approach integrates proximity-based pre-contact detection and VAC with QP-based motion constraints to proactively adjust impedance parameters while maintaining stable and controlled motion. By dynamically modulating stiffness and damping in response to sensor feedback, the system improves both obstacle avoidance performance and smooth contact handling. Additionally, a passivity-preserving energy tank mechanism mitigates instability arising from parameter variations, ensuring robust and adaptive behavior. Furthermore, experiments involving HRI$^{1}$ demonstrate that the proposed method ensures both safe avoidance and smooth contact handling. These results suggest that the proposed approach is highly applicable to safety-critical tasks in collaborative and industrial robotic environments.