Multiple Attacks Detection in Cyber-Physical Systems Using Random Finite Set Theory.

Area of Science:

• Cyber-physical systems security
• Network intrusion detection
• Sensor network security

Background:

• Cyber-physical systems (CPS) are vulnerable to simultaneous cyber attacks targeting multiple sensors.
• Existing detection methods often overlook the complexity of multiple, concurrent attacks.
• Efficiently detecting numerous attacks is crucial for maintaining CPS integrity.

Purpose of the Study:

• To develop an efficient method for detecting multiple cyber attacks on different sensors within CPS.
• To simultaneously determine the number of attacks and identify the specific sensors under attack.
• To address the gap in research concerning the detection of multiple cyber attacks in CPS.

Main Methods:

• Formulation of the multiple cyber attack detection problem using random finite set (RFS) theory.
• Application of an iterative RFS-based Bayesian filter for detection.
• Utilizing an approximation of the RFS-based Bayesian filter for computational efficiency.

Main Results:

• Demonstrated effectiveness of the RFS-based approach through four numerical experiments.
• Successful simultaneous detection of the number of attacks and the attacked sensors.
• Validation of the proposed method's capability in handling diverse attack scenarios.

Conclusions:

• The RFS-based approach provides an effective solution for detecting multiple cyber attacks in CPS.
• The method offers simultaneous detection of attack count and sensor compromise.
• This research advances the field of CPS security by addressing a critical, previously under-addressed problem.