Method for generating schedule information on basis of operation information, and server for implementing same
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional refrigerator control methods rely on predetermined temperature settings, leading to inefficient cooling or freezing and increased power consumption, as they do not adapt to actual usage patterns.
Innovation Solution
A method for generating schedule data for a refrigerator based on learned action patterns from multiple refrigerators, using a server to optimize prediction and control of compressor operations, thereby improving energy efficiency and reducing power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If the refrigerator operates based on predetermined temperature settings, then the temperature control is simple, but the power consumption increases and cooling efficiency decreases
Solution Approach 1:
The refrigerator performs preliminary learning of usage patterns during an initial period, storing action data and time data in advance. This pre-acquired knowledge enables the refrigerator to predict future door openings and adjust compressor operations proactively, reducing power consumption without requiring complex real-time decision-making systems.
Solution Approach 2:
The refrigerator implements a feedback mechanism where actual door opening actions are recorded as action data, compared with predicted actions, and used to continuously improve prediction accuracy. This closed-loop learning system refines the usage pattern model over time, enabling progressively better energy optimization without increasing control complexity.
2Measurement precision
If the refrigerator learns usage patterns individually, then the learning process is simple, but the prediction accuracy is insufficient
Solution Approach 1:
The refrigerator merges its individual usage pattern data with aggregated data from multiple other refrigerators. By combining action data and time data from a plurality of refrigerators, the system creates a more comprehensive and accurate prediction model, improving prediction accuracy while distributing the data processing burden.
3Measurement precision
If the refrigerator accumulates data over a long period, then the learning accuracy improves, but the time required for adaptation increases
Solution Approach 1:
The refrigerator performs preliminary data accumulation during a predetermined learning period, storing action data and time data in advance. This pre-acquired data foundation enables faster convergence to accurate predictions, reducing the overall adaptation time while maintaining high learning accuracy.
Solution Approach 2:
The refrigerator copies and utilizes aggregated usage pattern data from other refrigerators to supplement its own learning process. This copying mechanism allows the refrigerator to benefit from collective experience, achieving accurate predictions more quickly without requiring extensive individual data accumulation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed herein is a method for generating schedule data based on action data, and a server and a refrigerator implementing the same. The method for generating schedule data based on action data according to an embodiment of the present invention includes, in a refrigerator including one or more divided storage spaces, a step of a storage unit of the refrigerator storing pattern base data including action data performed by the refrigerator and time data, a step of a communication unit of the refrigerator receiving a first learning data set from a server, a step of a schedule generation unit of the refrigerator generating first schedule data including predicted action data of the refrigerator and time data of the predicted action data by mapping a first learning data set to the pattern base data, and a step of a control unit of the refrigerator controlling an action of the refrigerator based on the first schedule data.