Research Analyzer
← Back ICRA 2026

Autonomous UAV�Quadruped Docking in Complex Terrains Via Active Posture Alignment and Constraint-Aware Control

Haozhe Xu, Cheng Cheng, Hongrui Sang, Zhipeng Wang, Qiyong He, Xiuxian Li, Bin He

PDF

AI summary

Key figure (auto-extracted from paper)
A novel framework enables reliable autonomous UAV landing on dynamically moving quadruped robots in GPS-denied, rugged terrains by actively stabilizing the ground robot's posture and enforcing strict flight constraints.
UAV docking quadruped robot active posture alignment constraint-aware control deep reinforcement learning GPS-denied navigation

Problem

Existing docking systems target wheeled robots limited to flat terrain, while quadruped platforms exhibit unstable torso postures that prevent safe UAV landing in complex, GPS-denied environments.

Approach

The method combines a deep reinforcement learning policy that actively flattens the quadruped's torso during docking with a three-phase UAV controller that uses vision-based detection and constraint-aware sliding mode control to safely track and land on the moving platform.

Key results

  • HIM-HA reinforcement learning policy actively stabilizes quadruped torso posture for landing
  • Three-phase UAV docking strategy with FOV-constrained NFTSMC-BF controller
  • Real-world validation achieving successful docking on 17 cm stairs and 30-degree slopes
  • First demonstration of autonomous UAV-quadruped docking in unstructured 3D terrains

Why it matters

Enables robust aerial-ground collaboration for search, rescue, and exploration in GPS-denied, rugged environments where wheeled platforms cannot operate.

Abstract

Autonomous docking between Unmanned Aerial Vehicles (UAVs) and ground robots is essential for heteroge- neous systems, yet most existing approaches target wheeled plat- forms whose limited mobility constrains exploration in complex terrains. Quadruped robots offer superior adaptability but un- dergo frequent posture variations, making it difficult to provide a stable landing surface for UAVs. To address these challenges, we propose an autonomous UAV–quadruped docking frame- work for GPS-denied environments. On the quadruped side, a Hybrid Internal Model with Horizontal Alignment (HIM-HA), learned via deep reinforcement learning, actively stabilizes the torso to provide a level platform. On the UAV side, a three- phase strategy is adopted, consisting of long-range acquisition with a median-filtered YOLOv8 detector, close-range tracking with a constraint-aware controller that integrates a Nonsingular Fast Terminal Sliding Mode Controller (NFTSMC) and a logarithmic Barrier Function (BF) to guarantee finite-time error convergence under field-of-view (FOV) constraints, and terminal descent guided by a Safety Period (SP) mechanism that jointly verifies tracking accuracy and platform stability. The proposed framework is validated in both simulation and real-world scenarios, successfully achieving docking on outdoor staircases higher than 17 cm and rough slopes steeper than 30 degrees. Supplementary materials and videos are available at: https://uav-quadruped-docking.github.io.

Index terms

Cooperating Robots Legged Robots

Related papers