Agent- and Data-Driven IoT Security: From Analysis to Autonomous Protection

Dear Colleagues,

With the rapid proliferation of Internet of Things devices and the increasing complexity of IoT ecosystems, IoT security is facing significant challenges, including massive heterogeneous data generation, expanding attack surfaces, and the growing difficulty of ensuring end-to-end protection across distributed and resource-constrained environments. Traditional security analysis approaches—primarily based on rules, feature engineering, and task-specific models—are increasingly insufficient when dealing with multi-source heterogeneous data and evolving threats in IoT scenarios, revealing limitations in scalability, generalization, and adaptability.

Recent advances in artificial intelligence, particularly large language models, agentic AI paradigms, and other data-driven techniques, offer new opportunities to enhance IoT security. By enabling semantic understanding, contextual reasoning, cross-domain knowledge integration, and increasingly autonomous decision-making, these approaches can support more intelligent analysis of diverse data sources, including device logs, firmware, source code, network traffic, and threat intelligence. Such capabilities facilitate more effective anomaly detection, threat identification, and automated response, contributing to a shift from fragmented, tool-driven security solutions toward more integrated, intelligent, and partially autonomous defense mechanisms.

This Special Issue aims to bring together recent advances in IoT security, covering theoretical methods, system design, and real-world applications. Topics of interest include, but are not limited to:

• Security, privacy, and trust frameworks for agent-enabled IoT systems, supporting reliable and autonomous protection across distributed and resource-constrained environments;
• Robustness of agent-based IoT security mechanisms against adversarial threats, including evasion, poisoning, backdoor, and manipulation attacks in dynamic IoT ecosystems;
• Privacy-preserving techniques for data-driven IoT security, such as secure data sharing, federated learning, and lightweight privacy mechanisms for distributed devices;
• Resource-efficient learning, inference, and decision-making for IoT agents, considering constraints in computation, communication, and energy;
• Autonomous threat detection, response, and protection in IoT systems, enabling adaptive and self-evolving defense capabilities;
• Coordination, orchestration, and communication among distributed security agents, ensuring reliable collaboration under heterogeneous and dynamic network conditions;
• Detection and mitigation of malicious behaviors in IoT environments;
• Integration of agent-based and data-driven approaches with traditional IoT security mechanisms, such as intrusion detection, firmware analysis, and network monitoring;
• Secure deployment, update, and lifecycle management of IoT security models and agents, including trusted execution, secure updates, and resilience against tampering;
• Practical and scalable IoT security solutions for real-world applications, including smart homes, industrial IoT, smart cities, and critical infrastructures.

Prof. Dr. Yongle Chen
Dr. Wei Liu
Dr. Zhuangzhuang Zhang
Dr. Jianhua Wang
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Electronics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• IoT security
• agent-driven IoT security
• autonomous protection
• security, privacy, and trust
• adversarial robustness
• privacy-preserving mechanisms
• distributed security agents
• practical IoT security solutions