CPS Forecast Error Thresholding for Fast Anomaly Source Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current cyber-physical systems (CPS) face challenges in detecting anomalies in a timely manner, as existing methods are inefficient and often result in delayed detection, which can lead to safety issues and increased costs due to the complexity and costliness of traditional monitoring systems.

Innovation Solution

A system is developed that builds a CPS feature values forecasting model to calculate a total error threshold based on CPS characteristics, allowing for early anomaly detection by identifying when the total forecast error exceeds the calculated threshold, and pinpointing the source of the anomaly by determining the contribution of each feature to the total error.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional monitoring systems are used to detect anomalies in CPS, then detection can be performed, but the time to detect anomalies is delayed and system complexity increases

Engineering Contradiction:
Improveanomaly detection capabilityVSAvoidtime to detect anomaly
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary actions by building a forecasting model during normal operation to learn the expected behavior patterns of CPS features. This model continuously generates forecasts even before anomalies occur, enabling immediate detection when actual values deviate from predictions, thus reducing anomaly detection time without sacrificing reliability

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements feedback by continuously comparing actual CPS feature values against forecasted values and using the error analysis to identify anomalies. The feedback loop calculates total forecast errors, determines contribution ratios of individual features, and triggers alerts when anomalies are detected, creating a responsive monitoring mechanism that reduces detection delay

Inventive Principle:
Principle #23Feedback

2Reliability

If traditional monitoring systems are used to detect anomalies in CPS, then detection can be performed, but system complexity and cost increase

Engineering Contradiction:
Improveanomaly detection capabilityVSAvoidmonitoring system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system applies self-service by utilizing the CPS's own operational data and feature values to train and operate the forecasting model. The model learns from the system's normal behavior and monitors itself for anomalies, eliminating the need for separate complex monitoring infrastructure and reducing overall system complexity while maintaining detection reliability

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system changes parameters by transforming raw CPS feature values into forecasted values and error metrics. By monitoring changes in forecast errors rather than raw values directly, the system simplifies the monitoring task and reduces complexity while maintaining reliable anomaly detection capability

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If forecasting model is used to detect anomalies, then detection accuracy improves, but false responses may occur

Engineering Contradiction:
Improveanomaly detection accuracyVSAvoidfalse responses
Core Design Contradiction:
Measurement precisionVSObject-generated harmful factors

Solution Approach 1:

The system applies partial action by not treating all forecast errors as anomalies. Instead, it calculates the total forecast error and compares it against a dynamically determined threshold, triggering anomaly detection only when errors exceed this threshold. This selective approach improves detection accuracy while reducing false responses by filtering out normal variations

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3674946B1System and method for detecting anomalies in cyber-physical system with determined characteristics
Publication Date: 2021.08.11 AO KASPERSKY LAB
  • EP3674946B1 patent drawingFigure 1a
  • EP3674946B1 patent drawingFigure 1b
  • EP3674946B1 patent drawingFigure 1c

AI summary

Systems and methods for determining a source of anomaly in a cyber-physical system (CPS). A forecasting tool can obtain a plurality of CPS feature values during an input window and forecast the plurality of CPS feature values for a forecast window. An anomaly identification tool can determine a total forecast error for the plurality of CPS features in the forecast window, identify an anomaly in the cyber-physical system when the total forecast error exceeds a total error threshold, and identify at least one CPS feature as the source of the anomaly.