An Adaptive Protection System for Sensor Networks Based on Analysis of Neighboring Nodes.

Area of Science:

• Computer Science
• Cybersecurity
• Network Engineering

Background:

• Sensor systems are crucial for industrial, military, consumer, and agricultural applications, offering economic and operational benefits.
• The increasing reliance on sensor networks amplifies cybersecurity risks, necessitating robust attack detection mechanisms.
• Internal threats pose a significant challenge to the integrity and security of interconnected sensor systems.

Purpose of the Study:

• To develop and present an effective anomaly detection method for sensor networks.
• To enhance the collective security of sensor networks against internal cyberattacks.
• To enable timely detection and differentiation of various attack types within sensor systems.

Main Methods:

• A decentralized anomaly detection approach where sensor nodes monitor their neighbors' behavior.
• Utilizing common node parameters such as network traffic, Central Processing Unit (CPU) load, and packet drop rates for monitoring.
• Implementing a collective defense strategy where nodes collaborate to identify deviations.

Main Results:

• Demonstrated the ability to detect anomalies indicative of internal attacks within sensor networks.
• Successfully distinguished between different types of cyberattacks based on collected sensor node data.
• Validated the effectiveness of the proposed method through experimental results and a defined dataset collection strategy.

Conclusions:

• The proposed neighbor-observing anomaly detection method provides a viable solution for securing sensor networks against internal threats.
• Collective monitoring and analysis of intrinsic node parameters enable robust and scalable cybersecurity for sensor systems.
• This approach contributes to the overall resilience and trustworthiness of smart city, smart farm, and other IoT applications.