Semantic BIM Localization for Robots on Dynamic Building Sites

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Robots deployed in dynamically changing environments face challenges with localization and navigation due to reliance on non-permanent objects, leading to inaccurate maps and compromised navigation, requiring manual operator intervention for updates.

Innovation Solution

Utilizing a semantic model, such as a Building Information Model (BIM), to provide precise information on permanent objects and their schedules, enabling robots to autonomously update localization maps and avoid non-permanent objects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If robots rely on non-permanent objects for localization and navigation, then initial map creation is simpler, but localization accuracy deteriorates in dynamic environments

Engineering Contradiction:
ImproveInitial map creation simplicityVSAvoidLocalization accuracy
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent segments localization reference points into two categories: permanent objects (walls, columns, fixed structures) and non-permanent objects (movable equipment, temporary structures). By selectively using only permanent objects as localization references, the system maintains high localization accuracy in dynamic environments while still allowing simple initial map creation using all detectable features.

Inventive Principle:
Principle #1Segmentation

2Productivity

If robots use traditional SLAM algorithms for map creation, then initial localization is achieved, but manual operator intervention is required for updates in dynamic environments

Engineering Contradiction:
ImproveInitial mapping efficiencyVSAvoidMap update automation
Core Design Contradiction:
ProductivityVSExtent of automation

Solution Approach 1:

The patent implements a feedback mechanism where the robot continuously compares current sensor data against the BIM model during traversal. When discrepancies are detected (new permanent objects or changes), the system automatically updates the localization map. This closed-loop feedback eliminates the need for manual operator intervention while maintaining high mapping accuracy.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent performs preliminary actions by pre-acquiring BIM data containing semantic information about permanent objects before the robot begins traversal. This advance preparation allows the robot to immediately compare real-time sensor data with expected permanent object locations, enabling automated detection and update of environmental changes without manual intervention.

Inventive Principle:
Principle #10Preliminary action

3Adaptability or versatility

If robots traverse dynamic environments with changing objects, then operational flexibility is improved, but localization reliability deteriorates due to non-permanent objects

Engineering Contradiction:
ImproveOperational flexibilityVSAvoidLocalization reliability
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent applies local quality by using different object types for different purposes: permanent objects (walls, columns, fixed structures) are used exclusively for localization and navigation references, while non-permanent objects (movable equipment, temporary structures) are used for task execution but excluded from localization. This selective application maintains localization reliability while allowing operational flexibility in dynamic environments.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4285199B1Semantic models for robot autonomy on dynamic sites
Publication Date: 2026.03.04 BOSTON DYNAMICS INC
  • EP4285199B1 patent drawingFigure 1A
  • EP4285199B1 patent drawingFigure 1B
  • EP4285199B1 patent drawingFigure 2A

AI summary

A method (300) includes receiving, while a robot (100) traverses a building environment (10), sensor data (134) captured by sensors (132, 132a-n) of the robot. The method includes receiving a building information model (BIM) (30) for the environment that includes semantic information (32) identifying permanent objects (PO) within the environment. The method includes generating localization candidates (212, 212a-n) for a localization map (202) of the environment. Each localization candidate (212) corresponds to a feature of the environment identified by the sensor data and represents a potential localization reference point (222). The localization map is configured to localize the robot within the environment. For each localization candidate, the method includes determining whether the respective feature corresponding to the respective localization candidate is a permanent object (PO) in the environment and generating the respective localization candidate as a localization reference point (222) in the localization map for the robot.