State-Specific Threshold Calculation for Home IoT Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing anomaly detection methods for IoT devices in home networks fail to accurately detect anomalies due to variations in communication usage caused by changes in facility conditions, such as the number of residents or their activities, leading to detection failures or erroneous detections.
Innovation Solution
A threshold value calculation device and method that considers device states and occupancy states, including power states, setpoint values, and communication logs, to calculate tailored threshold values for accurate anomaly detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single threshold value is used for anomaly detection, then the detection process is simple, but detection accuracy deteriorates when communication usage varies due to facility conditions
Solution Approach 1:
The patent segments the anomaly detection system by creating multiple threshold values corresponding to different occupancy states (e.g., occupied, unoccupied, partially occupied). Instead of using a single threshold for all conditions, the system divides the detection space into multiple segments, each with its own optimized threshold, thereby maintaining high detection accuracy across varying facility conditions while keeping the overall process manageable through automated state-based selection
Solution Approach 2:
The patent implements dynamic threshold adjustment by linking threshold values to real-time occupancy states. The threshold is no longer static but dynamically changes based on the current facility condition (occupied/unoccupied states). This dynamic approach allows the system to adapt to varying communication usage patterns caused by different occupancy conditions, significantly improving anomaly detection accuracy without requiring manual reconfiguration
2Measurement precision
If threshold values are calculated for each combination of device states and occupancy states, then anomaly detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing multiple threshold values corresponding to different occupancy states and device states before actual anomaly detection begins. This pre-computation approach eliminates the need for complex real-time calculations during operation, reducing computational burden and system complexity while maintaining high detection accuracy. The system simply selects the appropriate pre-calculated threshold based on current state conditions
Solution Approach 2:
The system implements self-service through automated threshold selection based on detected occupancy states. Instead of requiring manual configuration or complex real-time optimization, the system automatically determines the current occupancy state and selects the corresponding threshold value from pre-calculated sets. This self-service mechanism simplifies operation and reduces system complexity while maintaining high detection accuracy across varying conditions
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A threshold value calculation device calculates a threshold value used by an anomaly detection device that detects anomalous communication of a first device provided in a facility in which a home network (11) is installed. In the facility, a second device that is different from the first device is provided. The threshold value calculation device includes: a device state acquirer (110) that acquires a device state of the first device during a first period; an occupancy state determiner (120) that determines an occupancy state by people in the facility during the first period based on information acquired from the second device; a communication log collector (130) that collects a communication log generated based on communication transmitted and received by the first device during the first period; and a learner (140) that calculates the threshold value for the communication transmitted and received by the first device during a second period that is a period after the first period based on the device state, the occupancy state, and the communication log. The device state includes one or more states of the first device. The occupancy state includes one or more states of the people. The learner (140) calculates the threshold value for each of combinations of the one or more states of the first device and the one or more states of the people.