In-home-presence probability calculation method, server apparatus, and in-home-presence probability calculation system
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting in-home presence are limited in accuracy, as they rely on power usage patterns over extended periods, which do not account for individual behavior variations and do not effectively utilize device operation patterns to determine if a person is at home at specific times.
Innovation Solution
An in-home-presence probability calculation method that uses device operation data to create histograms of time differences between device usage and expected leaving times, allowing for more accurate predictions of a person's presence based on past behavior patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If power usage patterns over extended periods are used to predict in-home presence, then a baseline prediction method is established, but accuracy is limited due to not accounting for individual behavior variations
Solution Approach 1:
The patent segments the prediction system into multiple components: power usage analysis, device operation pattern analysis, and correlation-based prediction. Each component processes specific types of data independently, then combines results to improve overall accuracy while maintaining manageable complexity through modular design
Solution Approach 2:
The patent changes the parameters used for prediction from only power usage metrics to include device operation times and their correlations with absence times. This parameter expansion allows the system to capture individual behavior variations while using established statistical methods to manage complexity
2Measurement precision
If device operation patterns and correlation analysis are added to improve prediction accuracy, then individual behavior variations are captured, but system complexity increases
Solution Approach 1:
The device management server performs multiple functions: collecting power usage data, collecting device operation data, analyzing correlations, and generating predictions. This multi-functionality consolidates complexity into a single server rather than requiring separate systems for each function
Solution Approach 2:
The patent creates histograms that copy and represent complex behavioral patterns in a simplified visual format. These histograms serve as simplified models of the underlying data, making it easier to analyze correlations between device operations and absence times without processing the raw data directly
3Productivity
If accurate in-home presence prediction is achieved, then delivery operations become more efficient, but more data collection and processing is required
Solution Approach 1:
The patent extracts only the most relevant features from the collected data: operation times of devices and their correlations with absence times. Rather than processing all available data, the system extracts specific temporal patterns that directly impact prediction accuracy, reducing the effective data volume while maintaining productivity benefits
Data Source
AI summary
A management method manages devices in a home includes, receiving operation information on one of the devices when an operation is performed for the one of the devices, specifying times at which the operation was performed, in accordance with the information, and receiving, whenever a state is entered when no one is expected to be home, time information when the state has been entered. Specifying in-home-absence times at which a state has been entered when no one is expected to be home, according to the time information, for each of the devices, calculating time differences for the specified operation times, and for each of the devices, specifying the performed operation as a first behavior of the person correlated to a behavior before the person leaves home, when variation of the calculated time differences of the corresponding one of the devices is equal to or lower than a threshold.