Adaptive Planning Framework for UAV-Based Surface Inspection in Partially Unknown Indoor Environments

TL;DR

The paper tackles UAV-based indoor surface inspection under partial knowledge by introducing a hierarchical planning framework that couples a segment-based global viewpoint planner with an inspection-oriented local adaptive planner. The global planner segments the reference map, generates informative viewpoints, and optimizes their sequence, while the local planner employs a static B-spline trajectory, MPC-based collision avoidance, and an occlusion-aware view-angle adaptation to maintain coverage amid unknown obstacles. Key contributions include the segment-based global coverage strategy, the integration of dynamic obstacle avoidance with view-angle adaptation, and validated robustness through simulations and real tunnel flights. The approach enhances inspection efficiency and safety in environments where reference maps are incomplete or outdated, with practical impact for industrial facilities and infrastructure maintenance.

Abstract

Inspecting indoor environments such as tunnels, industrial facilities, and construction sites is essential for infrastructure monitoring and maintenance. While manual inspection in these environments is often time-consuming and potentially hazardous, Unmanned Aerial Vehicles (UAVs) can improve efficiency by autonomously handling inspection tasks. Such inspection tasks usually rely on reference maps for coverage planning. However, in industrial applications, only the floor plans are typically available. The unforeseen obstacles not included in the floor plans will result in outdated reference maps and inefficient or unsafe inspection trajectories. In this work, we propose an adaptive inspection framework that integrates global coverage planning with local reactive adaptation to improve the coverage and efficiency of UAV-based inspection in partially unknown indoor environments. Experimental results in structured indoor scenarios demonstrate the effectiveness of the proposed approach in inspection efficiency and achieving high coverage rates with adaptive obstacle handling, highlighting its potential for enhancing the efficiency of indoor facility inspection.

Figures (4)

• Figure 1: A UAV performing autonomous inspection in a tunnel-like indoor environment using the proposed method.
• Figure 2: System overview of the proposed adaptive inspection framework. The planning module takes the reference map, localization, and perception information as input and generates a collision-free trajectory as the output. The planning module includes a global planner for viewpoint generation and sequencing and a local planner for B-spline reference trajectory generation, MPC-based tracking and collision avoidance, and adaptive view angle adjustment to ensure coverage and safe navigation during execution.
• Figure 3: Simulation results in two indoor environments for evaluating inspection coverage. 1a), 2a): Top-down views of the simulated environments with static and dynamic obstacles. Blue arrows show the movement directions of the dynamic obstacles. 1b), 2b): Planned global paths on the reference maps. 1c), 2c): Voxel grids show the coverage results of camera observations in static scenes (without unforeseen static and dynamic obstacles). 1d), 2d): Voxel grids show the coverage results in the environments shown in 1a) and 2a) and are visualized over the environment.
• Figure 4: Real-world inspection experiment in a tunnel environment. a) The physical test environment with multiple static obstacles. b) The reference map and the planned global paths. c) Voxel grids show the coverage results in the real flight.