Distributed Anomaly Detection Control for Resource-Limited Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing control systems face challenges in securing sufficient resources for anomaly detection, making it difficult to implement predictive maintenance effectively due to high resource requirements for monitoring phenomena in a shorter cycle.
Innovation Solution
A control system with multiple processing resources that allows for flexible arrangement of state value collection and anomaly detection means, enabling efficient anomaly detection by distributing these functions across available processing resources and using a support device to determine optimal resource allocation based on factors like state values, processing resource specifications, and network load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If monitoring is performed in a shorter cycle to detect anomalies earlier, then anomaly detection capability is improved, but processing resource requirements increase
Solution Approach 1:
The patent segments the anomaly detection system into multiple independent processing resources (first processing resource for state value collection, second processing resource for anomaly detection). This segmentation allows the system to distribute the computational load across multiple units, enabling shorter monitoring cycles without overwhelming a single processor, thus improving anomaly detection capability while managing resource requirements.
Solution Approach 2:
The patent creates processing resources that can be universally applied to different anomaly detection scenarios. The first processing resource collects state values that can represent various types of phenomena (vibration, temperature, pressure), and the second processing resource performs anomaly detection on diverse state values. This multi-functionality allows the same processing architecture to handle different monitoring requirements without proportionally increasing resource consumption.
2Reliability
If anomaly detection is implemented in existing control systems, then predictive maintenance is improved, but resource constraints are worsened
Solution Approach 1:
The patent divides the control system into distinct functional segments: existing control functions remain in original processing resources, while anomaly detection functions are separated into dedicated processing resources. This segmentation allows predictive maintenance to be implemented without complicating the existing control logic, as the anomaly detection module operates independently with its own resource allocation.
Solution Approach 2:
The patent introduces an intermediary mechanism where the first processing resource acts as a mediator between the control target and the second processing resource. It collects and prepares state values, then transmits them to the anomaly detection resource. This intermediary layer simplifies integration into existing systems by providing a standardized interface, reducing the complexity of resource management.
Data Source
AI summary
A control system controls a control target, and includes plural processing resources available for execution of arithmetic processing. The processing modules include a collection module and an anomaly detection module. The collection module collects one or more state values corresponding to a detection target included in the control target. The anomaly detection module calculates a value indicating a possibility that an anomaly has occurred in the detection target based on a feature value calculated from the one more state values having been collected. Each of the collection module and the anomaly detection module is capable of being arranged in a processing resource among the plural processing resources.