An Evidence Theoretic Approach for Traffic Signal Intrusion Detection.

Area of Science:

• Intelligent Transportation Systems
• Cybersecurity in Infrastructure

Background:

• Current traffic signal Intrusion Detection Systems (IDSs) primarily rely on connected vehicle data and image analysis, limiting their effectiveness against attacks targeting in-road sensors, controllers, and signals.
• Attacks on traffic signals are increasing globally, highlighting the urgent need for more robust detection mechanisms.

Purpose of the Study:

• To propose and validate an enhanced IDS for traffic signals that detects anomalies in traffic flow rate, phase time, and vehicle speed.
• To extend previous work by incorporating additional traffic parameters and advanced statistical tools for improved intrusion detection.

Main Methods:

• A novel IDS was developed, theoretically modeled using Dempster-Shafer decision theory and Shannon's entropy to quantify observation uncertainty.
• A simulation model was created using the SUMO traffic simulator, incorporating real-world scenarios and data from the Victorian Transportation Authority.
• Attack scenarios included jamming, Sybil attacks, and false data injection to test the system's efficacy.

Main Results:

• The proposed IDS achieved an overall detection accuracy of 79.3% in detecting various simulated attacks.
• The system demonstrated a reduction in false alarm rates compared to existing methods.

Conclusions:

• The developed IDS effectively detects intrusions by analyzing traffic flow anomalies, offering a significant improvement over existing systems.
• This research contributes to enhancing the security and reliability of intelligent traffic signal management systems against sophisticated cyber threats.