Assistant Task Offloading Using Processing-Rate Network Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems rely solely on signal strength metrics for determining whether to offload assistant-related processing tasks, leading to inefficiencies and processing delays due to disparities in network hardware, particularly in vehicles with more powerful antennas, resulting in suboptimal data transmission and reception.
Innovation Solution
Implementing network metrics such as processing rate and data reception acknowledgment to determine whether to offload tasks to a server computing device, allowing the client device to make informed decisions based on the server's data processing capabilities and connection status, rather than solely on signal strength.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If signal strength metrics are used to determine whether to offload tasks to a server, then the decision-making process is simple and fast, but the accuracy of the decision is compromised due to hardware disparities between different devices
Solution Approach 1:
The patent changes the parameter used for decision-making from signal strength (dB) to processing rate (packets per second). This parameter transformation allows the system to make accurate assessments of connection status that are independent of antenna hardware disparities, as processing rate directly reflects the server's ability to handle data rather than the client's transmission capability
Solution Approach 2:
The patent introduces an intermediary measurement approach where the client device measures the server's processing rate by sending test packets and timing the responses. This intermediary metric serves as a reliable indicator of connection status that mediates between the client's perspective and the server's actual processing capacity, eliminating the need to account for different antenna capabilities
2Reliability
If vehicle antennas with higher signal strength are used, then network connections can be established more easily, but processing delays occur because the application incorrectly assumes optimal connection conditions
Solution Approach 1:
The patent implements a feedback mechanism where the client device continuously monitors the server's processing rate and uses this feedback to dynamically adjust its offloading decisions. When the processing rate falls below a threshold, the system switches to local processing, preventing the time loss that would occur from attempting to offload tasks over a degraded connection
Solution Approach 2:
The patent makes the offloading decision dynamic by continuously measuring processing rate and adjusting behavior in real-time. Rather than relying on static signal strength measurements that don't reflect current server capacity, the system adapts its processing strategy based on the measured processing rate, switching between offloading and local processing as conditions change
3Productivity
If data offloading to server is attempted based on signal strength thresholds, then network resources are utilized, but unnecessary resource usage occurs when server processing capacity is insufficient
Solution Approach 1:
The patent applies partial action by sending only test packets to measure processing rate before committing to full data offloading. This preliminary measurement action allows the system to avoid the excessive resource consumption of attempting to offload large amounts of data when the server's processing capacity is insufficient, while still utilizing network resources effectively when capacity is adequate
Data Source
AI summary
Implementations set forth herein relate to off-loading, or temporarily ceasing such off-loading, computational tasks to a separate computing device based on a network metric(s) that is not limited to signal strength. Rather, a network metric for determining whether to continue relying on a network connection with a server computing device for certain computational tasks can be based on a current, or recent, interaction with the server computing device. In this way, an application executing at a computing device having a powerful antenna—but an otherwise limited network velocity, can determine to temporarily rely exclusively on local processing. For instance, an automated assistant can temporarily cease communicating audio data to a remote server computing device, during a dialog session, in response to determining a network metric fails to satisfy a threshold—even though there may appear to be adequate signal strength to effectively transmit the audio data.


