Cell-Based AR Positioning Grid for Millimeter Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current augmented reality (AR) systems face limitations in positioning accuracy due to the use of graphics processing units (GPUs) optimized for 32-bit single-precision calculations, which is insufficient for large-scale applications, leading to fluctuations and misalignment of AR objects in the real world.
Innovation Solution
A cell-based system that uses a reference grid with cells defined by single-precision floating-point coordinates, allowing millimeter-level precision positioning of AR content by associating each object with a local cell within the grid, enabling efficient identification and rendering of nearby objects without complex distance calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If 32-bit single-precision coordinates are used for positioning AR objects, then rendering performance is maintained, but positioning precision deteriorates to about 1 in 2^23 resolution
Solution Approach 1:
The patent divides the large-scale geographic space into multiple local cells, each with its own local coordinate system. AR objects are positioned within these local cells using 32-bit coordinates, which maintains rendering performance while achieving high positioning precision within each cell. The cell structure allows the system to overcome the limited range of 32-bit coordinates by providing a hierarchical organization of space.
Solution Approach 2:
The patent introduces a hierarchical dimension to the coordinate system by combining local cell coordinates with cell identification. Instead of using a single global coordinate system, the system adds a layer of organization where each location is defined by both its local coordinates within a cell and the cell's identity, enabling planet-scale coverage while maintaining 32-bit precision within each local context.
2Measurement precision
If 64-bit double-precision coordinates are used for positioning, then positioning precision is improved, but rendering performance deteriorates to unacceptably low speeds
Solution Approach 1:
The patent segments the coordinate representation into two parts: cell identification (which can use 64-bit precision if needed) and local coordinates within the cell (which use 32-bit single-precision). This segmentation allows the system to achieve high positioning precision for cell identification while maintaining fast rendering performance for local coordinate transformations and AR object positioning within each cell.
Solution Approach 2:
The patent applies different precision requirements to different aspects of positioning. Local coordinates within each cell use 32-bit precision optimized for GPU rendering performance, while cell identification can use higher precision when needed. This local quality approach ensures that computational performance is maintained where it matters most (in local rendering) while still achieving accurate global positioning through the cell hierarchy.
3Area of stationary object
If a global coordinate system is used for large-scale AR applications, then coverage area is increased, but positioning precision deteriorates due to 32-bit resolution limits
Solution Approach 1:
The patent divides the global coverage area into multiple local cells, each with its own local coordinate system that maintains 32-bit precision. This segmentation allows the system to cover planet-scale areas while ensuring that each local cell maintains high positioning precision suitable for AR applications. The cell-based approach effectively resets the precision reference point for each local region.
Solution Approach 2:
The patent adds a hierarchical dimension to the coordinate system by organizing space into cells with unique identifiers. This allows the system to achieve both global coverage and local precision by combining cell-level identification with local coordinate positioning, effectively extending the usable range of 32-bit coordinates through the cell hierarchy structure.
4Measurement precision
If higher bit-precision is used for coordinates, then positioning accuracy is improved, but storage requirements and bandwidth usage increase
Solution Approach 1:
The patent segments coordinate data into cell identification and local coordinates, allowing the system to store only the necessary precision for each component. Local coordinates within cells use compact 32-bit representations, while cell identification can be efficiently indexed. This segmentation reduces overall storage requirements compared to using high-precision coordinates for the entire global space.
Solution Approach 2:
The patent applies precision efficiently by using 32-bit coordinates for local positioning within cells, which provides sufficient accuracy for AR applications. This local quality approach avoids the need to store high-precision coordinates globally, reducing storage requirements and bandwidth usage while maintaining the positioning accuracy needed for the application.
Data Source
AI summary
Cell-based augmented reality (AR) content positioning systems may include a reference grid of cells, each of which includes a 32-bit intracellular coordinate system based on a respective reference point of the cell. Cell topology is selected such that the intracellular coordinate systems may utilize single-precision floating point numbers while retaining the ability to define content positions with, e.g., millimeter-level precision. Accordingly, rendering of AR content may be performed at a high precision using 32-bit devices and methods.