GPP-BIM — Global path planning for robot navigation using building information models

Mobile robots are increasingly being adopted for automated building inspections to improve safety, inspection accuracy, and operational efficiency. Global path planning (GPP) is required for automated robotic inspections and usually performed on prerecorded maps generated by robots. In the construction industry, building information modeling (BIM) is ubiquitous, and BIM models can be used for GPP, eliminating the need for prerecording maps and leveraging geometric and semantic information stored in BIM models. Due to their inherent complexity, BIM models are not directly applicable to state-of-the-art path planning algorithms. Typically, navigation models that filter and structure the vast amount of information relevant to navigation are derived from BIM models to perform GPP. Current research addressing the GPP problem using BIM models is mostly performed across single floors, may provide suboptimal paths, may suffer from long computation times, and may not consider the size of robots for collision checking. This paper presents a two-level GPP framework, termed “GPP-BIM” that leverages BIM models for enhanced robot navigation. Level 1 uses BIM to generate topological maps, while level 2 creates occupancy maps for each room. The two-level approach combines the efficiency and semantic integration of topological map planning with the detailed geometric considerations of occupancy maps, such as accommodating robot dimensions and detecting obstacles. The GPP-BIM framework is implemented in the Robot Operating System (ROS) 2 and uses the Industry Foundation Classes (IFC) open standard of BIM models. The GPP-BIM framework is validated on multiple IFC-based BIM models. The results of the validation tests show that GPP-BIM enables fast planning of obstacle-free and near-optimal paths in BIM environments. The GPP-BIM code is available as open-source software.

Subjects

Building information modeling (BIM)

Global path planning

Industry Foundation Classes (IFC)

Occupancy grid maps

Robot Operating System (ROS)

Topological maps

DDC Class

629.892: Robot

624.1: Structural Engineering

006.3: Artificial Intelligence

Lizenz

https://creativecommons.org/licenses/by/4.0/

Publication version

publishedVersion

Name

1-s2.0-S147403462600501X-main.pdf

Type

Main Article

Size

3.17 MB

Format

Adobe PDF

Options

GPP-BIM — Global path planning for robot navigation using building information models