Tethered UAV Autonomous Knotting on Environmental Structures for Transport

摘要

Cable-driven robotic systems are widely adopted for transport tasks due to their high load-bearing efficiency. However, their deployment in unstructured or unknown environments is hindered by the challenge of rapidly and reliably anchoring the cable endpoint. This work introduces a deployable cable-driven transport system that combines a tethered unmanned aerial vehicle (UAV) with a winch mechanism to autonomously form a topologically stable entanglement for cable anchoring. At the core of the system is a modular knot planner that integrates human-in-the-loop enclosing plane extraction, frontier-based enclosing path search, and knotting trajectory generation, incorporating metrics such as enclosing planarity, tether visibility, and tether clearance. In real-world experiments conducted in an urbanized outdoor environment, the system autonomously interpreted high-level user commands, executed a full knotting operation around a target structure, and successfully lifted a 15.3-kg payload to a height of 3.5 m. Beyond real-world trials, simulation studies confirmed the system's shape-agnostic knotting capability. A set of ablation experiments further demonstrated the necessity and effectiveness of these joint optimization metrics. Together, these results highlight the practicality and robustness of the proposed system for autonomous heavy-load transport in complex and previously unprepared environments, offering new capabilities for rapidly deployable robotic logistics.