Server Terminal State Estimation Using Pre-Generated Filters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for managing and estimating the state of multiple terminals using time series data are inefficient due to high processing loads and time requirements, especially when detecting abnormal states, as they involve converting data into feature values and searching databases across multiple terminals.
Innovation Solution
A system where a common search model is distributed to each terminal to generate filters, which are then collected by a server to estimate terminal states, reducing the need for data transmission and processing by using these filters to determine terminal states without querying each terminal directly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data is converted into feature values and searched across multiple terminals using existing methods, then the state of terminals can be detected, but the processing load and time requirements become excessively high
Solution Approach 1:
The patent applies preliminary action by pre-generating filters at each terminal based on historical time series data before any detection query is made. These filters are then uploaded to the server in advance, so when state detection is needed, the server can directly compare target data against pre-existing filters without performing complex conversions or searches across multiple terminals, thus resolving the contradiction between detection accuracy and processing efficiency
Solution Approach 2:
The patent uses copying by creating filter representations of terminal data characteristics. Instead of storing or processing actual time series data across multiple terminals, the system generates compact filter copies that capture essential state information. These filter copies can be quickly compared against target data to determine terminal states, maintaining detection accuracy while dramatically reducing processing load
2Measurement precision
If data is converted into feature values and searched across multiple terminals, then terminal states can be identified, but the time required for detection increases significantly
Solution Approach 1:
The system performs preliminary action by pre-computing filters from historical data and storing them on the server before detection is needed. When state identification is required, the comparison between target data and pre-existing filters occurs immediately without time-consuming data conversion or multi-terminal searching, thus reducing detection time while maintaining identification accuracy
Solution Approach 2:
The patent extracts essential state information from time series data by converting it into compact filter representations. This extraction process creates simplified data structures that capture key characteristics needed for state identification, allowing rapid comparison and significantly reducing the time required for detection while preserving the accuracy needed for correct state identification
3Measurement precision
If extensive data transmission is performed across multiple terminals for state detection, then accurate state estimation can be achieved, but the processing load increases
Solution Approach 1:
The patent uses copying by creating filter representations that encapsulate terminal data characteristics. Instead of transmitting actual time series data across the network, the system transmits compact filter copies that contain sufficient information for accurate state estimation. This approach maintains measurement precision while dramatically reducing processing load and data transmission requirements
Solution Approach 2:
The system extracts essential state information from raw time series data by generating filters that capture key characteristics. This extraction eliminates unnecessary data transmission and processing, as only the essential filter representations need to be handled, thus reducing device complexity and processing load while preserving state estimation accuracy
Data Source
AI summary
In an information processing device, a provision means provides multiple terminals with a common search model. A filter acquisition means acquires, from each terminal, a filter configured by a feature value obtained using the search model. A data acquisition means acquires, from each terminal, a target time series data subjected to state estimation. A estimation means estimates a state of the terminal which outputted the target time series data, based on the target time series data and the filter.