TY - GEN
T1 - Multi-agent Framework with Blockchain-Based Audit Trails for Securing Industrial Greenhouse Gas Monitoring Infrastructure
AU - Abiodun, Timileyin
AU - Nwulu, Nnamdi
AU - Olukanmi, Peter
N1 - Publisher Copyright:
© IFIP International Federation for Information Processing 2027.
PY - 2027
Y1 - 2027
N2 - Accurate and trustworthy Greenhouse Gas (GHG) monitoring systems are essential for emission accountability mechanisms, which support the accomplishment of global net-zero emission goals. Recent studies reveal urban inventories underreport emissions by 18.3%, and methane emissions in the energy sector are 70% higher than official figures. This study proposes iGuard, a novel framework combining multi-agent autonomous Artificial Intelligence (AI) with distributed ledger technology to create a vulnerability-resistant, real-time emission monitoring infrastructure. The agents employ a Long Short-Term Memory (LSTM)-based autoencoder for temporal anomaly detection, a random forest model for identifying network intrusions in Internet of Things (IoT) communication channels, and cryptographic hashing for verification. All agents operate under a consensus mechanism to provide unified threat assessments, with security events immutably recorded on the blockchain for forensic audit trails. Evaluation on the EPA CEMS and NF-ToN-IoT datasets achieved 100% tampering detection, 95.61% network intrusion accuracy, and consensus-driven precision improvements of up to 5.78 percentage points, while maintaining sub-500 ms latency.
AB - Accurate and trustworthy Greenhouse Gas (GHG) monitoring systems are essential for emission accountability mechanisms, which support the accomplishment of global net-zero emission goals. Recent studies reveal urban inventories underreport emissions by 18.3%, and methane emissions in the energy sector are 70% higher than official figures. This study proposes iGuard, a novel framework combining multi-agent autonomous Artificial Intelligence (AI) with distributed ledger technology to create a vulnerability-resistant, real-time emission monitoring infrastructure. The agents employ a Long Short-Term Memory (LSTM)-based autoencoder for temporal anomaly detection, a random forest model for identifying network intrusions in Internet of Things (IoT) communication channels, and cryptographic hashing for verification. All agents operate under a consensus mechanism to provide unified threat assessments, with security events immutably recorded on the blockchain for forensic audit trails. Evaluation on the EPA CEMS and NF-ToN-IoT datasets achieved 100% tampering detection, 95.61% network intrusion accuracy, and consensus-driven precision improvements of up to 5.78 percentage points, while maintaining sub-500 ms latency.
KW - Blockchain
KW - Cyber-physical Systems
KW - Greenhouse Gas Monitoring
KW - Industrial IoT Security
KW - Multi-Agent System
UR - https://www.scopus.com/pages/publications/105042873953
U2 - 10.1007/978-3-032-28939-1_32
DO - 10.1007/978-3-032-28939-1_32
M3 - Conference contribution
AN - SCOPUS:105042873953
SN - 9783032289384
T3 - IFIP Advances in Information and Communication Technology
SP - 471
EP - 491
BT - Technological Innovation to Tackle Societal Challenges - 17th IFIP WG 5.5 Doctoral Conference on Computing, Electrical and Industrial Systems, DoCEIS 2026, Proceedings
A2 - Camarinha-Matos, Luis M.
A2 - Ferrada, Filipa
PB - Springer Science and Business Media Deutschland GmbH
T2 - 17th Advanced Doctoral Conference on Computing, Electrical and Industrial Systems, DoCEIS 2026
Y2 - 17 June 2026 through 19 June 2026
ER -