Slow drift aware dynamic risk assessment in cyber physical systems using quantum neutrosophic fuzzy modelling

Area of Science:

• Cyber-Physical Systems Security
• Network Intrusion Detection
• Risk Assessment Methodologies

Background:

• Cyber-physical systems (CPS) face reliability threats from slow degradation in network behavior, characterized by gradual shifts in traffic patterns.
• Existing research often overlooks these subtle, long-term network variations, leaving CPS vulnerable to faults, stealthy intrusions, and aging.
• This gap necessitates advanced methods to analyze and mitigate slow degradation impacting CPS security.

Purpose of the Study:

• To propose a novel Quantum State-based Exponentially Weighted Moving Average (QS-EWMA) approach for analyzing slow drift in CPS network behavior.
• To enhance the detection of stealthy intrusions and assess risks associated with network degradation in CPS, particularly in medical device contexts.
• To improve the overall reliability and security of cyber-physical systems through advanced dynamic risk assessment.

Main Methods:

• Implementation of QS-EWMA for slow drift analysis in network traffic characteristics.
• Integration of a Network Intrusion Detection System (NIDS) with Sparse Random Walk Softplus S-shaped Rectified Gated Recurrent Unit (SRWSSR-GRU) for intrusion detection.
• Application of dependency-aware aggregation using Choquet k-Additive Function Integral Model (CkAFIM) and uncertainty-based risk assessment via Min-Max Normalization-based Neutrosophic Logic System (MMN-NLS).
• Enhancement of explainability using SHapley Max-Abs Scaling Additive exPlanation (SMAS-HAP) and decision-making via Markov Linear Discrete-Time Propagation Decision Process (MLDTPDP).

Main Results:

• The proposed QS-EWMA-based system effectively identified slow degradation and intrusions in CPS network behavior.
• The system achieved a notable Indeterminacy Detection Rate (IDR) of 13.43%, demonstrating superior performance compared to existing methods.
• Enhanced explainability and robust decision-making were achieved through integrated advanced models.

Conclusions:

• The QS-EWMA approach offers a significant advancement in detecting subtle network degradations and intrusions within CPS.
• The comprehensive risk assessment framework, incorporating dependency-aware aggregation and uncertainty analysis, enhances CPS security.
• The proposed methodology provides a superior and explainable solution for dynamic risk assessment in cyber-physical systems, particularly for critical applications like medical devices.