Cluster Inc. Patents Automated 3D Space Generation Technology, Creating Interactive Environments from BIM/CAD Data in Minutes
Cluster Inc. announced it has obtained a patent (No. 7827282) on March 2, 2026, for an information processing technology that automatically converts real-world data into interactive 3D spaces. The new technology takes input data such as BIM, CAD, and PLATEAU, using AI to automatically identify objects and assign behaviors and textures. This reduces the generation time for a building-scale 3D space from several months to just tens of minutes, aiming to accelerate the use of digital twins in fields like construction, real estate, disaster prevention, and education.
📋 Article Processing Timeline
- 📰 Published: May 20, 2026 at 17:30
- 🔍 Collected: May 20, 2026 at 09:01
- 🤖 AI Analyzed: May 20, 2026 at 09:20 (18 min after Collected)
Cluster Inc. (Headquarters: Shinagawa-ku, Tokyo; CEO: Naoto Kato) announced that it has obtained a patent (No. 7827282), registered on March 2, 2026, for an information processing technology that automatically converts physical spaces into 3D environments.
## Patented Technology Overview
The patented technology generates 3D spatial data by taking inputs like BIM, CAD, PLATEAU 3D city models, LiDAR point clouds, and 3D Gaussian Splatting. An AI then automatically determines the type of objects and assigns corresponding behaviors, textures, and collision detection. The resulting 3D spaces are accessible on smartphones, tablets, PCs (via browser), and VR headsets such as Meta Quest.
## From Static to Interactive 3D
While technologies to display 3D spaces from BIM or PLATEAU data already exist, adding motion and textures required manual configuration for each object, a process that could take weeks to months and significant cost for a single building. This new technology automates this workflow, enabling the generation of a comparably scaled 3D space in about tens of minutes, according to the company's internal verification (processing time may vary depending on the scale and structure of the input data).
Specifically, the AI automatically assigns properties upon data input:
- **Doors**: Open and close when their handles are grabbed.
- **Walls, Floors, Obstacles**: Collision detection (colliders) makes them solid, causing users to stop or bounce off.
- **Materials**: Textures and visual properties (materials) like wood grain, gloss, or transparency are applied to surfaces like wood, metal, and glass.
- **Cityscapes**: For 3D city models like PLATEAU, appropriate textures and collision detection are applied to buildings, roads, and trees, allowing users to navigate the city on foot.
This enables the generation of 'enterable, touchable, and movable' 3D spaces, from a single building to an entire city, without manual setup.
## Core Technology
The technology features three distinct AI functions that work in sequence:
1. **Attribute Determination Function**: Identifies the type of object (e.g., door, vehicle, wall) based on its shape and identifiers (name, ID).
2. **Action Inference Function**: Infers the necessary actions for the identified object (e.g., opening, operating, colliding).
3. **Script Generation Function**: Automatically generates the program (script) to replicate the inferred actions within the 3D space.
This end-to-end process eliminates the need for individual object configuration, autonomously building an interactive 3D space from input data.
## Use Case Examples
- **Construction**: Instantly generate a walkable, finished building from BIM data for client presentations.
- **Factories/Warehouses**: Digitize existing facilities for safety training, workflow simulation, and simple layout planning.
- **Disaster Prevention**: Recreate real buildings for multi-participant evacuation drills.
- **Urban Planning**: Allow residents to experience a proposed redevelopment area in 3D to facilitate consensus-building.
- **Education/Training**: Digitize facilities like farms or factories for hands-on learning by multiple students.
- **Real Estate**: Enable immersive virtual property tours from a single floor plan.
- **Retail**: Digitize proposed store layouts to test shelf placement and customer flow.
## Future Development
Cluster plans to integrate this technology into its existing 3D space generation services and offer it on a subscription basis. The company will explore applications in a wider range of fields, including construction, disaster prevention, education, and real estate, to accelerate the social implementation of 'human-in-the-loop digital twins.'
## Patent Information
- Patent Number: 7827282
- Title of Invention: Information Processing Method, Information Processing Apparatus, Information Processing Program, and Information Processing System
- Registration Date: March 2, 2026
- Inventor: Cluster, Inc.
## Patented Technology Overview
The patented technology generates 3D spatial data by taking inputs like BIM, CAD, PLATEAU 3D city models, LiDAR point clouds, and 3D Gaussian Splatting. An AI then automatically determines the type of objects and assigns corresponding behaviors, textures, and collision detection. The resulting 3D spaces are accessible on smartphones, tablets, PCs (via browser), and VR headsets such as Meta Quest.
## From Static to Interactive 3D
While technologies to display 3D spaces from BIM or PLATEAU data already exist, adding motion and textures required manual configuration for each object, a process that could take weeks to months and significant cost for a single building. This new technology automates this workflow, enabling the generation of a comparably scaled 3D space in about tens of minutes, according to the company's internal verification (processing time may vary depending on the scale and structure of the input data).
Specifically, the AI automatically assigns properties upon data input:
- **Doors**: Open and close when their handles are grabbed.
- **Walls, Floors, Obstacles**: Collision detection (colliders) makes them solid, causing users to stop or bounce off.
- **Materials**: Textures and visual properties (materials) like wood grain, gloss, or transparency are applied to surfaces like wood, metal, and glass.
- **Cityscapes**: For 3D city models like PLATEAU, appropriate textures and collision detection are applied to buildings, roads, and trees, allowing users to navigate the city on foot.
This enables the generation of 'enterable, touchable, and movable' 3D spaces, from a single building to an entire city, without manual setup.
## Core Technology
The technology features three distinct AI functions that work in sequence:
1. **Attribute Determination Function**: Identifies the type of object (e.g., door, vehicle, wall) based on its shape and identifiers (name, ID).
2. **Action Inference Function**: Infers the necessary actions for the identified object (e.g., opening, operating, colliding).
3. **Script Generation Function**: Automatically generates the program (script) to replicate the inferred actions within the 3D space.
This end-to-end process eliminates the need for individual object configuration, autonomously building an interactive 3D space from input data.
## Use Case Examples
- **Construction**: Instantly generate a walkable, finished building from BIM data for client presentations.
- **Factories/Warehouses**: Digitize existing facilities for safety training, workflow simulation, and simple layout planning.
- **Disaster Prevention**: Recreate real buildings for multi-participant evacuation drills.
- **Urban Planning**: Allow residents to experience a proposed redevelopment area in 3D to facilitate consensus-building.
- **Education/Training**: Digitize facilities like farms or factories for hands-on learning by multiple students.
- **Real Estate**: Enable immersive virtual property tours from a single floor plan.
- **Retail**: Digitize proposed store layouts to test shelf placement and customer flow.
## Future Development
Cluster plans to integrate this technology into its existing 3D space generation services and offer it on a subscription basis. The company will explore applications in a wider range of fields, including construction, disaster prevention, education, and real estate, to accelerate the social implementation of 'human-in-the-loop digital twins.'
## Patent Information
- Patent Number: 7827282
- Title of Invention: Information Processing Method, Information Processing Apparatus, Information Processing Program, and Information Processing System
- Registration Date: March 2, 2026
- Inventor: Cluster, Inc.