Spatial Anchor Sharing for XR Scene Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing artificial reality (XR) devices face challenges in efficiently sharing and aligning scene data across multiple devices, leading to inconsistencies and increased processing requirements.
Innovation Solution
The implementation of spatial anchor sharing allows XR devices to retrieve and align spatial anchors, enabling accurate scene data sharing and alignment without the need for each device to rescan the environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If each XR device rescans the environment to gather scene data, then the device can independently map the environment, but the processing requirements and time increase significantly
Solution Approach 1:
The patent applies preliminary action by having the first XR device scan and gather scene data before the second device arrives. The scene data including spatial anchors is stored and can be retrieved by subsequent devices, eliminating the need for them to rescan the same environment.
Solution Approach 2:
The patent implements copying by allowing the second XR device to retrieve and use a copy of the scene data gathered by the first device. Instead of each device independently scanning, the scene representation is copied and shared between devices through spatial anchor data.
2Loss of information
If multiple XR devices independently gather scene data, then each device has complete scene information, but inconsistencies arise in virtual object placement across devices
Solution Approach 1:
The patent applies merging by combining the scene data from multiple devices through spatial anchor sharing. The coordinate frames and spatial anchors from different devices are merged into a unified scene representation, ensuring consistent virtual object placement across all devices.
Solution Approach 2:
The patent uses spatial anchors as an intermediary to mediate between different device coordinate systems. The anchors serve as reference points that allow transformation and alignment of scene data from different devices, ensuring consistency without requiring direct device-to-device coordination.
3Ease of operation
If XR devices share scene data through centralized storage, then data accessibility improves, but processing overhead and data transmission requirements increase
Solution Approach 1:
The patent extracts only the essential spatial anchor data from the complete scene data for sharing between devices. Instead of transmitting entire scene representations, only the critical spatial anchor information needed for coordinate frame alignment is extracted and shared, reducing data transmission overhead.
4Manufacturing precision
If XR devices align coordinate frames using spatial anchors, then scene consistency improves, but the complexity of coordinate transformation increases
Solution Approach 1:
The patent implements self-service by having each XR device independently perform coordinate frame alignment using the shared spatial anchor data. Each device autonomously transforms its coordinate frame to match the unified scene coordinate system without requiring complex inter-device communication or coordination protocols.
Data Source
AI summary
Implementations described herein can share scene data between multiple artificial reality (XR) devices, such as head-mounted displays (HMDs). A first XR device can retrieve stored spatial anchors associated with a scene, and align a coordinate frame with coordinate frames for a second XR device using the spatial anchors. The second XR device can have identified physical objects in the scene with corresponding object types defined in a scene lexicon (e.g., with a camera on the second XR device capturing images of physical objects and performing object recognition or through manual tagging of objects with object types), and store the object types and locations of the physical objects as scene data. The first XR device can obtain the scene data based on associations between the scene data and the spatial anchors and render virtual objects using the scene data, without having to rescan the scene for the physical objects.