SLAM Golden Reference Synchronization in Edge Cloud HMDs

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Distributed Simultaneous Localization and Mapping (SLAM) systems in edge cloud architectures face challenges in maintaining synchronization among client nodes, leading to inaccuracies and drift in SLAM maps due to incorrect landmarks.

Innovation Solution

The system identifies and selects 'golden references' among landmarks, compares instances across multiple HMDs, and adjusts maps based on latency and computation requirements, using edge or cloud servers to resolve discrepancies and maintain accurate mapping.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Power

If distributed SLAM is implemented across multiple HMDs in edge cloud architecture, then processing capability and mapping coverage are improved, but synchronization accuracy and landmark reliability deteriorate due to drift and inaccuracies

Engineering Contradiction:
Improveprocessing capabilityVSAvoidsynchronization accuracy
Core Design Contradiction:
PowerVSMeasurement precision

Solution Approach 1:

The patent introduces a dedicated synchronization module that acts as an intermediary between multiple HMDs and the edge server. This module receives SLAM data from multiple HMDs, performs centralized processing to identify and resolve landmark discrepancies, and redistributes corrected data. The intermediary handles the complex synchronization logic separately, allowing each HMD to maintain high processing capability while relying on the central module to ensure synchronization accuracy through coordinated landmark validation and drift correction.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements a feedback mechanism where the synchronization module continuously monitors landmark data from multiple HMDs, detects drift and inconsistencies, and generates correction signals that are fed back to the respective HMDs. This closed-loop feedback process enables real-time adjustment of SLAM parameters and landmark positions, maintaining measurement precision across the distributed system without compromising the processing power of individual nodes.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If golden references are selected and compared across multiple HMDs, then landmark accuracy is improved, but computational complexity and data processing time increase

Engineering Contradiction:
Improvelandmark accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts and isolates the computationally intensive golden reference comparison logic into a dedicated synchronization module that operates separately from the main SLAM processing pipeline of each HMD. By taking out this complex computation to a centralized or edge-based module, individual HMDs can maintain simpler, more efficient local processing while the extracted synchronization function handles the heavy computational load of comparing golden references across multiple devices, thereby managing overall system complexity.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The system applies partial action by selectively comparing only the golden reference landmarks (a subset of all detected features) across HMDs rather than performing exhaustive comparisons of all SLAM data. This partial approach focuses computational resources on the most critical synchronization points, achieving sufficient landmark accuracy without the excessive computational complexity that would result from complete data comparison across all devices.

Inventive Principle:
Principle #16Partial or excessive action

3Stability of the object's composition

If SLAM data is synchronized across distributed HMDs, then map consistency is improved, but communication overhead and network bandwidth consumption increase

Engineering Contradiction:
Improvemap consistencyVSAvoidnetwork bandwidth consumption
Core Design Contradiction:
Stability of the object's compositionVSLoss of energy

Solution Approach 1:

The patent segments the SLAM data synchronization process into distinct components: local feature detection at each HMD, extraction of critical golden reference landmarks, selective transmission of only necessary correction data to the edge server, and targeted distribution of updates back to relevant HMDs. This segmentation allows the system to achieve map consistency by synchronizing only the essential landmark information rather than transmitting complete SLAM datasets, thereby reducing network bandwidth consumption while maintaining stability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system applies local quality by allowing each HMD to independently perform local feature detection and initial SLAM processing, maintaining high-quality local mapping capabilities. Only the critical golden reference landmarks that require cross-device validation are subjected to centralized synchronization. This approach ensures map consistency for critical features through network synchronization while allowing non-critical local data to remain decentralized, minimizing overall communication overhead.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS10937192B2Resolving incorrect distributed simultaneous localization and mapping (SLAM) data in edge cloud architectures
Publication Date: 2021.03.02 DELL PROD LP
  • US10937192B2 patent drawing
  • US10937192B2 patent drawing
  • US10937192B2 patent drawing

AI summary

Embodiments of systems and methods for resolving incorrect distributed Simultaneous Localization and Mapping (SLAM) data in edge cloud architectures are described. In some embodiments, a method may include: receiving a first plurality of landmarks usable to produce a first map of a physical space for a first Head-Mounted Device (HMD) and selecting a first set of golden references; receiving, from an Information Handling System (IHS) coupled to a second HMD, a second set of golden references usable to produce a second map of the physical space for the second HMD; determining that a first instance of a given golden reference in the first set of golden references matches a second instance of the given golden reference in the second set of golden references; and in response to the first and second instances of the given golden reference being distant from each other by a threshold, produce a new first map.