SLAM-Map Fabric for Indoor 3D Mapping Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current GPS-based mapping technologies are limited in accuracy and speed, particularly indoors and underground, and struggle with classifying and identifying 3-D objects in real-time, which hampers their effectiveness in providing location identification and augmented reality services.
Innovation Solution
A server-based SLAM-Map Fabric system that utilizes Simultaneous Localization And Mapping (SLAM) technology to create a feature cloud model of the world by integrating digital contributions from remote devices, enabling accurate 3-D mapping and object identification using machine learning classifiers, and allowing for augmented reality experiences through virtual object placement and triggers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS-based mapping technologies are used, then location identification service is provided, but mapping accuracy and speed are limited especially indoors and underground
Solution Approach 1:
The patent introduces SLAM technology as an intermediary system that works alongside GPS. When GPS is unavailable or inaccurate (indoors, underground), the SLAM system takes over by using sensors like accelerometers, gyroscopes, and magnetometers to perform local positioning and mapping, thus bridging the gap in coverage and maintaining service reliability across all environments
Solution Approach 2:
The mapping system is designed to be universal by integrating multiple positioning approaches (GPS, SLAM, Wi-Fi positioning) into a single platform. This multi-functional system can automatically switch between different positioning methods based on environmental conditions, ensuring both high accuracy and reliability whether the user is outdoors with GPS signal, indoors without GPS, or in various other scenarios
2Productivity
If traditional mapping systems are used, then location identification is provided, but real-time object classification and identification capability is lacking
Solution Approach 1:
The system performs preliminary actions by continuously building and updating a feature cloud model of the environment in real-time as the mobile device moves. This pre-established 3D model with embedded object features enables immediate object classification and identification when needed, eliminating the delay of processing images in real-time and achieving both high speed and complete information retention
Solution Approach 2:
The patent replaces traditional image-based object recognition mechanisms with a SLAM-based feature cloud approach. Instead of analyzing images in real-time, the system uses pre-processed 3D spatial features and machine learning classifiers to identify objects, substituting the computationally intensive image processing mechanism with a more efficient spatial reasoning mechanism that achieves real-time performance
3Measurement precision
If SLAM technology is used to create feature cloud model, then mapping accuracy and speed are improved, but system complexity increases
Solution Approach 1:
The patent segments the complex SLAM system into distinct functional modules: sensor data acquisition module, feature extraction module, 3D model construction module, and object classification module. Each module handles a specific aspect of the process, making the overall complex system more manageable and maintainable while preserving the high mapping accuracy benefits of SLAM technology
Data Source
AI summary
A non-volatile digital storage medium storing instructions that, when executed by a server system, establish on the server system computer processes for implementing a method of providing a server-based feature cloud model of a realm, the method comprising: receiving by a server a series of digital contributions that collectively originate from a plurality of remote computing devices, each digital contribution characterizing a defined and distinct three-dimensional volume of the realm, wherein each distinct three-dimensional volume includes a portion that does not overlap any other three-dimensional volume; processing by the server of each one of the received digital contributions to associate such contribution with a global coordinate system; and storing and integrating each of the processed contributions in a realm model database so as to provide the feature cloud model of the realm stored on the server.