Sparse SLAM Anchor Graphs for Mixed Reality Sharing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mixed-reality systems face challenges in sharing SLAM coordinate systems efficiently due to high network bandwidth and memory requirements, particularly when dealing with large mixed-reality worldspaces, leading to network performance and memory issues.
Innovation Solution
The implementation of anchor graphs, where mixed-reality devices and servers generate anchor vertices and edges that include key frames and transforms, allowing for the efficient sharing of SLAM coordinate systems by storing and transmitting only relevant data, reducing the need for complete mapping of all elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large numbers of key frames are shared between different devices, then accurate registration of SLAM coordinate systems is achieved, but network bandwidth consumption increases and memory requirements at each device increase
Solution Approach 1:
The patent extracts only the essential elements needed for coordinate system registration: anchor vertices representing key poses and anchor edges representing transformations between them. Instead of transmitting complete key frame data, the system transmits a sparse graph structure containing only the critical anchor points and their interrelationships, dramatically reducing data volume while preserving registration accuracy.
Solution Approach 2:
The patent segments the continuous SLAM coordinate system into discrete anchor vertices and edges. By dividing the complete key frame sequence into selected anchor poses connected by transformation edges, the system creates a manageable sparse representation that can be efficiently transmitted and stored while maintaining the overall coordinate system structure.
2Adaptability or versatility
If complete SLAM coordinate systems are transmitted to multiple devices, then seamless shared experiences are enabled, but memory consumption at each device increases
Solution Approach 1:
The patent creates a simplified copy of the complete SLAM coordinate system in the form of a sparse anchor graph. Instead of storing complete key frame data at each device, the system stores a compact representation consisting of anchor vertices and their transformation relationships, enabling multiple devices to share the same coordinate system with minimal memory footprint.
3Measurement precision
If large amounts of SLAM data are transmitted over the network, then accurate coordinate system sharing is achieved, but network performance deteriorates due to constrained bandwidth
Solution Approach 1:
The patent extracts only the critical transformation data needed for coordinate system registration. By transmitting anchor vertices and their inter-vertex transformations rather than complete key frame sequences, the system reduces network payload size significantly while maintaining the precision required for accurate coordinate system sharing.
Solution Approach 2:
Instead of transmitting complete key frame data and deriving transformations at the receiving end, the patent inverts the approach by pre-calculating and transmitting only the essential anchor-to-anchor transformations. This reversal of the data transmission strategy reduces network bandwidth requirements while preserving registration accuracy.
Data Source
AI summary
Mixed-reality systems are provided for using anchor data structures, such as anchor graphs, within a mixed-reality environment. These systems utilize anchor components, such as anchor vertexes, that comprise at least one first key frame, a first mixed-reality element, and at least one first transform connecting the at least one first key frame to the first mixed-reality element. Anchor connecting components, such as anchor edges, comprise transformations that connect the anchor components (e.g., anchor vertexes).