Server Session Activation for Voice Command Latency Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing voice command processing systems experience latency issues due to inefficient session management, particularly in cloud-based spoken dialogue systems with a large number of clients, leading to increased server requirements and observable delays in long-running dialogues.

Innovation Solution

A method for early activation of sessions during voice command recording, utilizing a session activation signal to associate with a session ID, and activating either a prior available server or a free server by loading session data from a database, allowing for immediate processing of voice commands without significant latency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If session activation is performed after voice command recording in conventional systems, then server resources can be conserved by keeping sessions inactive, but latency increases and user experience deteriorates

Engineering Contradiction:
Improvelatency timeVSAvoidserver resource consumption
Core Design Contradiction:
Loss of timeVSUse of energy by moving object

Solution Approach 1:

The patent applies preliminary action by activating the session on the server during or prior to the voice command recording phase, rather than waiting until after recording. This early session activation ensures that when the voice command is transmitted and processed, the server is already ready to handle it immediately, thereby reducing latency time while maintaining efficient resource utilization through selective activation based on recording detection

Inventive Principle:
Principle #10Preliminary action

2Productivity

If multiple servers are deployed to handle concurrent voice commands, then system capacity and reliability improve, but system complexity and cost increase

Engineering Contradiction:
Improvesystem capacityVSAvoidserver system complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements dynamics by enabling flexible server assignment and session migration capabilities. Sessions can be dynamically activated on different servers based on availability, and the system can handle server failures by activating sessions on alternative servers. This dynamic approach allows the system to scale capacity as needed while maintaining manageable complexity through standardized session management protocols

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent applies universality by creating a session management framework that works across multiple servers with consistent behavior. The session activation signal and session ID association mechanism provide universal functionality that can be implemented on any server in the system, allowing seamless load distribution and failover without increasing operational complexity

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

3Speed

If session data is retained in memory for quick access, then processing speed improves, but memory consumption increases and economic efficiency decreases

Engineering Contradiction:
Improveprocessing speedVSAvoidmemory consumption
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The patent applies preliminary action by pre-loading session data into server memory when a session activation signal is detected during voice command recording. This ensures that when the actual voice command processing occurs, the session data is already in memory for immediate access, achieving fast processing speeds while only consuming memory when actually needed, thus optimizing the balance between speed and resource consumption

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10665233B2Method for operating a server system and for operating a recording device for recording a voice command; server system; recording device; and spoken dialogue system
Publication Date: 2020.05.26 ROBERT BOSCH GMBH
  • US10665233B2 patent drawing
  • US10665233B2 patent drawing
  • US10665233B2 patent drawing

AI summary

A method for operating a server system that includes a plurality of servers for processing a voice command recorded by a recording device connected, via an interface, to the server system includes, in response to the recording of the voice command, reading in a session activation signal from the recording device; checking if there is an association between the session activation signal and a session ID; if it is established that there is the association between the session activation signal and the session ID, ascertaining an availability of a prior server that previously processed a session assigned to the session ID; and activating the session on the prior server if it is available, but, if it is not available, then activating the session on a free server of the server system by loading the session from a session database. The voice command is received by the corresponding server during the activation of the session or even subsequent to the activation of the session, in order to be processed.