Hybrid Speech Recognition for Remote Controls

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing remote control devices face challenges in efficiently processing voice commands due to background noise, leading to incorrect command recognition and increased power consumption with the need for faster processors and more memory.

Innovation Solution

A method that distributes speech recognition performance between a remote control device and a voice platform in the cloud, using a combination of local and remote processing to enhance voice command recognition while reducing power, network, memory, and processing time usage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If voice input is continuously streamed to the cloud for processing, then speech recognition accuracy is improved, but network consumption increases

Engineering Contradiction:

Improvespeech recognition accuracyVSAvoidnetwork consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The speech processing system is segmented into two parts: local processing for trigger words and basic commands, and cloud processing for complex queries. This segmentation allows the device to send only necessary audio data to the cloud, reducing network consumption while maintaining recognition accuracy for cloud-handled queries.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Instead of continuously streaming all voice input to the cloud, the system performs partial processing locally first. Only when local processing determines cloud assistance is needed does it transmit audio, thus avoiding excessive network usage while still achieving accurate recognition when required.

Inventive Principle:

Principle #16Partial or excessive action

2Device complexity

If voice input is sent to the cloud for processing, then local processor requirements are reduced, but response time increases due to latency

Engineering Contradiction:

Improvelocal processor requirementsVSAvoidresponse time

Core Design Contradiction:

Device complexityVSLoss of time

Solution Approach 1:

Commands are segmented into two categories: those handled locally for immediate response and those requiring cloud processing. Trigger words and simple commands are processed locally without cloud communication, providing fast response times, while complex queries are routed to the cloud with appropriate latency tolerance.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary local processing of voice input to identify trigger words and simple commands before considering cloud communication. This preliminary action filters out many queries that don't require cloud processing, reducing overall latency by handling common cases locally.

Inventive Principle:

Principle #10Preliminary action

3Loss of time

If local speech recognition is performed, then response time is reduced, but power consumption and hardware requirements increase

Engineering Contradiction:

Improveresponse timeVSAvoidpower consumption

Core Design Contradiction:

Loss of timeVSUse of energy by moving object

Solution Approach 1:

Speech recognition tasks are segmented by complexity: trigger word detection and simple command recognition are performed locally with low power consumption, while complex natural language processing is delegated to the cloud, optimizing the balance between response time and power usage.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different processing qualities are applied locally vs. remotely. Local processing handles time-sensitive, low-complexity tasks with high responsiveness, while cloud processing handles complex tasks where latency is less critical, optimizing overall system power efficiency.

Inventive Principle:

Principle #3Local quality

4Device complexity

If a single digital assistant is configured in the remote control device, then device complexity is reduced, but adaptability to different task types decreases

Engineering Contradiction:

Improvedigital assistant configurationVSAvoidtask type coverage

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The remote control device is configured with a universal digital assistant that can interact with multiple specialized digital assistants through the cloud. This allows the device to maintain simple local configuration while achieving broad adaptability across different task types through cloud-based coordination of multiple specialized assistants.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentEP3676827B1Local and cloud speech recognition

Publication Date: 2025.03.05 ROKU INC

Figure 1
Figure 2
Figure 3

AI summary

Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for distributing the performance of speech recognition among a remote control device and a voice platform in the cloud. In some embodiments, the remote control device operates to receive a voice input from a user. The remote control device detects a trigger word in the voice input. The remote control device then processes the voice input. The remote control device then transmits the voice input to a voice platform based on the detecting in order to determine an intent associated with the voice input.