Safety-Aware Control for Human–Aerial Robot Cooperation in Environments With Obstacles

Abstract

This paper presents a safety-aware control framework for human-aerial robot co-transportation in environments with obstacles. A higher-order sliding mode observer is employed to estimate the relative velocity and human acceleration for human intention tracking, and the estimated states are incorporated into an impedance model to realize compliant following. To reduce collision risk in cluttered environments, a desired-force shaping method is proposed to proactively manipulate the transported object while simultaneously accounting for robot-obstacle avoidance. Possible conflicts between object-obstacle and robot-obstacle avoidance are explicitly addressed within a unified force-generation framework. Moreover, because the cable elasticity is unknown, large manipulation forces may cause excessive relative motion between the robot and the object. To improve safety, a barrier-Lyapunov-function-based variable stiffness strategy is introduced to keep the relative position within a prescribed safe range. Lyapunov-based analysis shows boundedness of the closed-loop signals and preservation of the relative-position constraint. The effectiveness of the proposed method is demonstrated by real-world experiments and simulation studies on collaborative transportation of a long bar in environments with obstacles. Note to Practitioners—This paper is motivated by human-robot co-transportation tasks in cluttered environments, where an aerial robot assists a human in carrying a long object through narrow spaces. In practice, passive compliance alone is often insufficient, since following human motion may still cause the object to collide with nearby obstacles, while the robot itself must also remain safe. To address this issue, this paper proposes a safety-aware control framework that integrates human intention estimation, compliant following, proactive object manipulation, and robot-obstacle avoidance. The robot normally follows the human through an impedance behavior, but actively adjusts the desired cable force when the object approaches obstacles. A variable stiffness strategy is further introduced to keep the relative robot-object position within a safe range. The method is relevant to applications such as indoor material handling and inspection support in constrained environments.