Dynamic Control Barrier Function-Based Obstacle Avoidance Planning of Robotic Arm for Safe Human–Robot Collaboration

Human–robot collaboration (HRC) requires robots to step out of enclosures and work in deep integration with human operators to achieve truly intelligent and flexible manufacturing. As a result, human safety has become a dominant issue, together with production efficiency. To ensure the safety of human operators while achieving efficient task completion, this paper proposes a dynamic obstacle avoidance planning framework that integrates visual perception and optimization control. The framework utilizes feedback from a depth-sensing system as input and relies on its core, the Control Barrier Function (CBF), to generate collision-free trajectory control for the entire robot body. To address the limitations of traditional CBF, we extend it to the Dynamic Control Barrier Function (DCBF) based on the HRC mode of speed and separation monitoring. Experiments conducted in both simulated and real-world scenarios validate the framework’s ability to enhance task efficiency while maintaining high safety standards, making it suitable for the HRC applications.