Beam Data Selection via Frequency Correlation for Wake-Up Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The high computing resource requirements and increased manufacturing costs of intelligent home devices due to parallel wake-up word detection across multiple beam data, which degrades user experience.

Innovation Solution

Selecting target beam data based on frequency sampling and correlation coefficients to prioritize processing on the most relevant data, reducing the need for comprehensive wake-up word detection across all beams.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If parallel wake-up word detection is performed on all beam data, then wake-up detection speed is improved, but processing resource consumption increases and manufacturing cost increases

Engineering Contradiction:
Improvewake-up detection speedVSAvoidprocessing resource requirements
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent segments the beam data processing into two stages: first performs wake-up detection on a small number of selected target beam data, then performs detection on remaining beam data only if needed. This segmentation reduces the parallel processing burden while maintaining detection speed by dividing the workload into priority-based batches.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by selecting specific target beam data with higher priority for wake-up detection based on beam quality metrics. Instead of uniform processing of all beams, it focuses computational resources on beams with better characteristics (higher signal-to-noise ratio, better correlation coefficients), thereby reducing overall processing requirements while maintaining detection effectiveness.

Inventive Principle:
Principle #3Local quality

2Speed

If parallel wake-up word detection is performed on all beam data, then wake-up detection speed is improved, but manufacturing cost increases

Engineering Contradiction:
Improvewake-up detection speedVSAvoidmanufacturing cost
Core Design Contradiction:
SpeedVSEase of manufacture

Solution Approach 1:

The patent segments wake-up detection into priority-based batches, performing detection first on target beam data with higher correlation coefficients and signal quality. This segmentation allows the system to achieve effective wake-up detection with fewer parallel processing units, reducing hardware complexity and manufacturing cost while maintaining speed.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the processing parameter by selectively applying wake-up detection to only the most promising beam data based on pre-calculated correlation coefficients and signal-to-noise ratios. This parameter-based selection reduces the number of parallel detection channels needed, thereby lowering manufacturing cost while preserving detection speed for critical cases.

Inventive Principle:
Principle #35Parameter changes

3Device complexity

If target beam data is selected based on correlation coefficients, then processing resource demands are reduced, but detection accuracy may be affected

Engineering Contradiction:
Improveprocessing resource demandsVSAvoiddetection accuracy
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent performs preliminary actions by calculating correlation coefficients and signal-to-noise ratios for all beam data before wake-up detection. This preliminary assessment identifies target beam data with highest likelihood of containing wake-up words, allowing the system to focus processing resources on high-probability candidates and maintain detection accuracy while reducing overall resource demands.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback by using correlation coefficients and signal quality metrics to guide the selection of target beam data. The system continuously evaluates beam characteristics and adjusts selection criteria based on observed performance, ensuring that processing resources are allocated to beams most likely to yield successful wake-up detection, thereby maintaining accuracy with reduced resources.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3657497B1Method and device for selecting target beam data from a plurality of beams
Publication Date: 2024.01.10 BEIJING XIAOMI INTELLIGENT TECH CO LTD
  • EP3657497B1 patent drawingFigure 1~2a
  • EP3657497B1 patent drawingFigure 2b
  • EP3657497B1 patent drawingFigure 2c

AI summary

The present invention relates to a method and device for selecting target beam data from a plurality of beams. The method includes: obtaining a plurality of beam data, and performing frequency sampling on each of the plurality of beam data; obtaining a plurality of beam frequency correlation coefficients based on frequency sampling data of each of the plurality of beam data, in which a beam frequency correlation coefficient is configured to indicate a similarity between one in the plurality of beam data and another one in the plurality of beam data; obtaining a beam frequency correlation coefficient sum corresponding to each of the plurality of beam data based on the plurality of beam frequency correlation coefficients; and selecting beam data having the beam frequency correlation coefficient sum satisfying a preset correlation coefficient requirement in the plurality of beam data as target beam data.