Active Search for Sensor Querying in Overlapping Edge Networks

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In geographically overlapping networks, identifying devices that are likely to have relevant data for a specific query is challenging due to the high cost of querying and the partial observability of data.

Innovation Solution

The system represents devices as nodes in a graph, associates each node with feature sets or feature vectors describing its capabilities, and uses Active Search algorithms to identify target nodes within a limited query budget, thereby optimizing the number of queries while maximizing the number of relevant nodes discovered.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If queries are performed to identify devices with relevant data in geographically overlapping networks, then the completeness of data collection is improved, but the query cost and time consumption increase significantly

Engineering Contradiction:

Improvecompleteness of data collectionVSAvoidquery time consumption

Core Design Contradiction:

Loss of informationVSLoss of time

Solution Approach 1:

The system pre-computes and stores feature vectors for each device node in the graph, including attributes like device type, location, and capabilities. This preliminary preparation allows the active search algorithm to quickly evaluate nodes without performing expensive queries, thereby reducing query time while maintaining data collection completeness

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces the mechanical approach of systematically querying all devices with an information-theoretic active search algorithm. This algorithm uses feature vectors and entropy-based scoring to intelligently select which nodes to query, substituting brute-force mechanical querying with a smarter information-driven approach that reduces time consumption

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Loss of information

Engineering Contradiction:

Improvecompleteness of data collectionVSAvoidquery cost

Core Design Contradiction:

Loss of informationVSLoss of energy

Solution Approach 1:

The active search algorithm performs partial action by selecting only a subset of nodes to query based on their feature vectors and calculated scores. Instead of querying all devices in the geographically overlapping networks, the system identifies and queries only the most promising nodes, thereby reducing query cost while still achieving sufficient data collection completeness

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system changes parameters by using feature vectors to represent devices and employing entropy-based scoring to prioritize nodes. This parameter transformation allows the system to efficiently identify high-value nodes for querying, reducing overall query cost while maintaining effective data collection

Inventive Principle:

Principle #35Parameter changes

3Reliability

If a large number of nodes are queried to ensure finding target nodes with desired capabilities, then the reliability of data acquisition is improved, but the complexity of the search process increases

Engineering Contradiction:

Improvereliability of data acquisitionVSAvoidcomplexity of the search process

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The active search algorithm incorporates feedback by using the results from queried nodes to update the search strategy. After each query, the system learns from the outcomes and adjusts its selection of subsequent nodes to query, improving reliability of data acquisition while managing search process complexity through adaptive learning

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs preliminary actions by pre-computing feature vectors for all nodes before the search begins. This preparation work includes organizing device attributes and capabilities in advance, which simplifies the actual search process and reduces the complexity of evaluating nodes during the query phase while maintaining reliable data acquisition

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20250045625A1Active search-based approach for sensor querying in geographically overlapping edge networks

Publication Date: 2025.02.06 DELL PROD LP

AI summary

An active search based approach for performing queries in networks, including geographically overlapping networks, is disclosed. After generating a graph representing devices operating in one or more networks, feature sets of the devices are retrieved and stored in corresponding nodes. When performing a query, a small set of nodes is used to train a model, such as a classifier, and the graph is searched for nodes that are part of a particular class. When a sufficient number of nodes are identified, which is much less than the number of nodes in the graph, the corresponding devices are queried and the resulting data may be used to perform an action.