Chenglong Fu

About

I study how connected systems and AI behave under adversarial conditions. My work examines device semantics, timing, and execution context, and develops analysis and verification methods for secure automation.

• IoT & CPS security: device capabilities, timing attacks, and cross-platform automation. Representative work: HAWatcher (USENIX Security 2021), Hidden Attributes (ACM CCS 2025), and IoT-Handler (RAID 2026).
• AI security & privacy: language-model jailbreaks, backdoor defenses, and federated learning. Recent work: Activation Dilution (ACM CCS 2026), UDPU (RAID 2026), and Does Efficiency Leak? (Inscrypt 2026).
• AI-assisted security analysis: program analysis, formal verification, and physical simulation. Related work: PG-DSL (ACSAC 2026), L2R (Inscrypt 2026), LLM-GridEval (EnergySP 2026, Workshop), and ICOAudit (IEEE/IFIP DSN 2026, Industry Track).

I co-direct the SmartHome Lab, am a member of CESAR, and lead the LLM Core Facility project at UNC Charlotte. I received my Ph.D. from Temple University, advised by Xiaojiang Du, and my bachelor's degree in Information Security from USTC.

News

• Oct. 2026Presented CESAR's research at RSTCON 2026 and joined the “Break and Hold the (Power)Line” panel on October 24.
• Oct. 2026Our papers received the RAID Best Paper Award and Inscrypt Best Paper Award (Security Track). [Certificates]
• Sep. 2026New papers accepted at ACM CCS, ACSAC, and Inscrypt. [Publications]
• Sep. 2026Released RKA Core 3.0.1 and EdgeLoom 0.2.0.
• Jun. 2026Served as a Poster Chair for DSN 2026 in Charlotte.

Selected peer-reviewed publications and accepted papers. Preprints are listed separately.

Showing 17 publications.

No matching publications.

Open-source Software

RKA — Research Knowledge Agent

A knowledge system for long-running, AI-assisted research. It connects research notes, experiments, decisions, and claims with supporting evidence, helping researchers retain context across sessions and coordinate AI collaborators while maintaining control over scientific decisions. [Code] [Documentation] [Paper]

EdgeLoom

A toolchain for inspecting and transforming smart-home edge drivers. It supports exposing hidden device capabilities, generating SmartThings proxy artifacts from Home Assistant entities, validating artifact structure, and preserving reviewable records of changes. [Code] [Documentation]

Selected Preprints

1. Research Should Outlive the Agent: RKA as Researcher-controlled Project Memory for Agentic Research

   Chenglong Fu. SSRN preprint, 2026.

2. Where Does Detection Fail? A Residual-Evidence Diagnostic Workflow for Industrial Control Systems

   Peiran Shi, Jian Xiang, Xiang Zhang, Chenglong Fu. arXiv:2608.02821v2, 2026.

Selected Awards

View certificate

Best Paper Award, RAID 2026

Effective Handling of Erroneous Executions of Automation Rules on Multiple Platform Smart Home IoT Systems.

[Paper] [Certificate PDF]

View certificate

Best Paper Award (Security Track), Inscrypt 2026

Does Efficiency Leak? Auditing the Privacy of Compressed and Stale Updates in Federated Intrusion Detection for IoT.

[Paper] [Certificate PDF]

CCI Summer Education Project award, 2026. Awarded $7,000 in project support to develop an entry-level cybersecurity course for students across all UNC Charlotte majors. [Details]

Activities & Professional Service

• Program Committee: IEEE Symposium on Security and Privacy 2027; USENIX Security 2027; IEEE/IFIP DSN 2026 (Research Track); ACM CCS 2025.
• Poster Chair: IEEE/IFIP DSN 2026.
• University service: University Research Council, 2026–2028; Research and Infrastructure Committee, Department of Cybersecurity, 2026; Graduate Committee, Department of Cybersecurity, 2026–2027.
• Research funding review: External Reviewer, Hong Kong Research Grants Council, Research Fellow Scheme, 2026/27.
• Talks & panels: RSTCON 2026, October 24 — presented “The Center for Energy Security and Reliability at UNC Charlotte” and joined the “Break and Hold the (Power)Line” panel.
• Research community: Co-director, SmartHome Lab; member, CESAR; PI, UNC Charlotte LLM Core Facility project.

Teaching & Curriculum Development

• ITIS 4221 / 5221: Penetration Testing and Secure Programming.
• ITIS 3200: Introduction to Security and Privacy.
• Summer 2026 curriculum project: Entry-Level Cybersecurity Course for All UNC Charlotte Majors. Supported by a $7,000 CCI Summer Education Project award.

Prospective Students & Collaborators

I welcome conversations about IoT/CPS security, AI security and privacy, verified automation, and open research infrastructure. Students interested in these areas are welcome to email a brief introduction, relevant experience, and the research questions they would like to explore.

Email: chenglong.fu@charlotte.edu