As the complexity of interior spaces grows, from hospitals and airports to high-density mixed-use developments, the need for accurate, efficient indoor surveying has become critical. Traditional terrestrial laser scanning (TLS) methods, while highly accurate, often require multiple scan positions and significant setup time. By contrast, Simultaneous Localization and Mapping (SLAM) enables mobile data capture that’s well-suited to GPS-denied interiors and can significantly shorten field time in corridor- and room-based buildings.
What Is SLAM?
SLAM is a class of algorithms that allows a device, typically using LiDAR and/or visual-inertial sensors, to build a map of an unknown environment while simultaneously estimating its own pose within that map.
In public safety and built-environment contexts, SLAM is central to broader location-based services initiatives focused on indoor mapping, tracking, and navigation.
The NIST Public Safety Communications Research (PSCR) program specifically supports real-time collaborative signal and inertial SLAM techniques for indoor mapping in first-responder use cases, where GPS is unavailable or unreliable.
SLAM vs. Traditional Laser Scanning
Controlled comparisons show that while Traditional Laser Scanning (TLS) still leads in absolute accuracy, the performance gap has narrowed.
In a 2025 study using long indoor corridors, static laser scanners such as the Leica RTC360 and Trimble X7 achieved root mean square errors (RMSE) between 1.2 mm and 5.4 mm, whereas newer-generation SLAM systems yielded 7.3 mm to 8.8 mm. Older SLAM devices performed significantly worse (~36 mm RMSE).
The study also highlights contexts in which SLAM errors grow with trajectory length, exemplified by a ~0.1 m drift over 150 m, underscoring cumulative error (drift) in long, loop-closure-poor environments.
Practically, however, SLAM’s mobility offers substantial time savings, particularly in interiors where repeated tripod redeployment for static scanners becomes a logistical bottleneck.
To evaluate SLAM systems under real-world conditions, the Hilti SLAM Challenge Dataset provides office, lab, and construction sequences with millimeter-level sparse ground truth.
The dataset doesn’t imply all SLAM systems achieve millimeter accuracy, but rather offers a high-precision benchmark against which their performance can be rigorously tested, especially in sparse, reflective, or feature-poor indoor environments.
Applications in AEC
In the architecture, engineering, and construction (AEC) sector, SLAM is increasingly seen as a practical tool for improving efficiency and reducing costs in indoor surveying workflows. It is particularly valuable for:
- Existing conditions capture: SLAM devices are now commonly used in scan-to-BIM workflows, enabling rapid as-built documentation for structures lacking accurate floor plans or historical records. Walk-through SLAM scanning can generate complete interior point clouds in a fraction of the time required by TLS setups.
- Renovation and retrofit planning: For buildings undergoing phased renovations or tenant fit-outs, SLAM scanning provides a fast, repeatable method for capturing current conditions, even in partially occupied or constrained spaces. Its mobility allows survey teams to work around daily operations with minimal disruption.
- Construction progress tracking: SLAM-equipped handheld or backpack-mounted systems are being deployed on active job sites to create regular 3D snapshots of project status. These data support schedule verification, trade coordination, and integration into 4D models. When fused with project control networks or GCPs, SLAM data can help align field conditions with design intent.
- Asset and facilities management: SLAM-generated models can be tied to asset inventories and MEP systems in facility management platforms. For owners and operators, this provides a dynamic, spatially accurate reference to support maintenance, space utilization, and lifecycle planning.
These applications are particularly well-suited to plan-dominant or corridor-heavy structures—such as healthcare, education, and commercial buildings, where a single SLAM scan path can replace dozens of static TLS setups, with minimal compromise in geometric accuracy for most AEC needs
Technical Challenges and Considerations
SLAM accuracy can degrade in environments with:
- Repetitive geometry (e.g., long identical corridors),
- Low lighting or visual texture (for visual SLAM),
- Highly reflective or transparent surfaces, and
- Extended trajectories without loop closures, which lead to drift.
The Hilti dataset is designed to challenge SLAM algorithms with precisely these conditions, making it a valuable resource for system developers and evaluators.
Hybrid SLAM approaches that incorporate sensor fusion, typically combining LiDAR, IMUs, and sometimes signal-based positioning, are showing strong potential.
These methods are supported by earlier government research into 3D indoor mapping for mobile robots, which addressed similar issues like occlusions, overhangs, and irregular architectural geometry.
