CBPM Sensor Attack Detection Using Bayesian and LSTM Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern US Navy ships and submarines with advanced sensor networks are vulnerable to cyber-attacks, particularly false data injection attacks (FDIAs), which can compromise condition-based predictive maintenance (CBPM) systems, leading to undetected failures and potential catastrophic consequences.
Innovation Solution
A cyber resilience integrated security inspection system (CRISIS) using a combination of deep learning algorithms (LSTM and GRU) and a graphical Bayesian network model to detect and mitigate FDIAs by analyzing sensor data and historical information, integrated with a software-in-the-loop and hardware-in-the-loop system for real-time monitoring and maintenance recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep learning algorithms (LSTM and GRU) are used to detect FDIAs, then detection accuracy is improved, but computational complexity and power consumption increase
Solution Approach 1:
The system segments the detection task by using multiple specialized deep learning models (LSTM for temporal patterns, GRU for gate-controlled sequences) rather than a single monolithic model. Each model processes specific aspects of sensor data, improving detection accuracy while distributing computational load to manage power consumption in resource-constrained edge devices
Solution Approach 2:
The system changes parameters by adjusting the architecture and configuration of deep learning models to optimize the balance between detection accuracy and power consumption. This includes tuning hyperparameters, selecting appropriate model depths, and adapting processing frequency based on operational conditions to reduce energy usage while maintaining high detection rates
2Reliability
If a hybrid software-in-the-loop and hardware-in-the-loop system is used for real-time detection, then system reliability is improved, but device complexity increases
Solution Approach 1:
The system merges software-in-the-loop simulation with hardware-in-the-loop real-time processing into a unified hybrid architecture. This integration allows virtual models and physical sensors to work together, improving reliability through cross-validation while managing complexity through unified control and shared data structures
Solution Approach 2:
The system introduces intermediary layers that facilitate communication and data exchange between the software simulation environment and hardware sensing components. These intermediaries manage data flow, synchronize timing, and coordinate operations, thereby improving system reliability while abstracting and managing the inherent complexity of the hybrid architecture
Data Source
AI summary
A method for detecting false data injection attacks (FDIAs) on a condition-based predictive maintenance (CBPM) system includes: collecting sensor data from sensors monitoring components of a system maintained by the CBPM system to extract features for a cyberattack detection model and gathering historical data of the system to build a cyberattack knowledge base about the system; combining the sensor data and the historical data to train the cyberattack detection model; using a graphical Bayesian network model to capture domain knowledge and condition-symptom relationships between the sensor-monitored components and the sensors; and based on the cyberattack detection model and the Bayesian network model, detecting the FDIAs on the CBPM system.