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-home times, allowing for more accurate predictions of a person's presence based on past behavior patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If power usage patterns over extended periods are used to predict in-home presence, then prediction coverage is improved, but prediction accuracy deteriorates due to not accounting for individual behavior variations
Solution Approach 1:
The patent segments the prediction approach by creating separate histograms for different devices (TV, refrigerator, air conditioner, etc.) and different time periods (weekday daytime, weekday nighttime, weekend daytime, weekend nighttime). This segmentation allows the system to capture individual behavior variations for each device and time period while maintaining comprehensive coverage across all devices and times.
Solution Approach 2:
The patent adds the dimension of device-specific operation patterns by creating histograms that track time differences between device operations and expected leaving-home times. This transforms the prediction from a single power usage metric to a multi-dimensional analysis incorporating device type, operation timing, and behavioral correlations, thereby improving accuracy while maintaining coverage.
2Measurement precision
If device operation data is collected and analyzed to improve prediction accuracy, then prediction accuracy is improved, but system complexity increases
Solution Approach 1:
The system automatically collects device operation data, creates histograms, and performs correlation analysis without requiring manual configuration or intervention. The device management server autonomously processes the data from multiple devices, calculates time differences, builds histograms, and generates presence predictions, thereby improving accuracy while keeping the user-side complexity low.
Solution Approach 2:
The device management server acts as an intermediary that centralizes the complex data processing and histogram creation operations. Instead of requiring complex interactions between multiple devices and systems, the server consolidates the analysis function, simplifying the overall system architecture while enabling sophisticated prediction accuracy through centralized data processing.
Data Source
AI summary
A management method manages a plurality of devices provided in a structure, a server includes storage storing operation information related to a plurality of devices and time information of a time at which it is expected that no one is present. The method specifies, for each device, operation times at which an operation is performed, operation information, specifying in-structure-absence times when it is expected that no one is present, both in accordance with the time information. For each device, time differences for the specified operation times are calculated. For each device, a performed certain operation is a behavior pattern correlated to a behavior when the person continues being present when a ratio of a frequency of the calculated time differences regarding the corresponding device, which is lower than a particular value, is equal to or lower than an other value.