User-Configurable Speech Recognition for Aircraft

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing aircraft speech recognition systems face challenges in balancing accuracy and latency, which affects user experience, particularly in time-sensitive operations, as they are often constrained by design-time tradeoffs and lack user-configurability.

Innovation Solution

A user-configurable speech recognition system that allows pilots to dynamically adjust the tradeoff between recognition accuracy and latency using a slider interface, selecting from various speech recognition model configurations tailored to individual preferences and operational contexts, utilizing acoustic and language models that can vary based on user-defined settings and specific aircraft phases.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If speech recognition accuracy is increased to improve user experience, then recognition precision improves, but latency increases which degrades user experience

Engineering Contradiction:
Improvespeech recognition accuracyVSAvoidlatency
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system dynamically adjusts speech recognition model configurations based on real-time operational context and user preferences. Multiple model configurations are maintained with different accuracy-latency tradeoffs, and the system selects appropriate configurations dynamically rather than using a fixed design-time setting, resolving the contradiction by making the system adaptable to changing requirements.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes key parameters of the speech recognition model including vocabulary size, grammar complexity, and processing depth based on operational context. By adjusting these parameters, the system can optimize for either accuracy or latency depending on the situation, resolving the fixed tradeoff contradiction through parameter variability.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If multiple speech recognition model configurations are provided for different accuracy-latency tradeoffs, then user configurability and adaptability improve, but system complexity increases

Engineering Contradiction:
Improveuser configurabilityVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The speech recognition system is segmented into multiple discrete model configurations, each optimized for specific accuracy-latency requirements. This segmentation allows the system to offer varied performance characteristics without requiring a single complex adaptive model, managing complexity through structured division into manageable configuration options.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system implements a universal architecture that can accommodate multiple model configurations through a common interface and selection mechanism. This multi-functionality allows the same system to serve different accuracy-latency requirements without proportionally increasing complexity, as the underlying infrastructure remains shared across all configurations.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP4095853B1Dynamic speech recognition methods and systems with user-configurable performance
Publication Date: 2023.11.15 HONEYWELL INTERNATIONAL INC
  • EP4095853B1 patent drawingFigure 1
  • EP4095853B1 patent drawingFigure 2
  • EP4095853B1 patent drawingFigure 3

AI summary

Methods and systems are provided for assisting operation of a vehicle using speech recognition. One method involves identifying a user-configured speech recognition performance setting value selected from among a plurality of speech recognition performance setting values, selecting a speech recognition model configuration corresponding to the user-configured speech recognition performance setting value from among a plurality of speech recognition model configurations, where each speech recognition model configuration of the plurality of speech recognition model configurations corresponds to a respective one of the plurality of speech recognition performance setting values, and recognizing an audio input as an input state using the speech recognition model configuration corresponding to the user-configured speech recognition performance setting value.