Unified Random Forest and Hybrid Bat Optimization Based Man-in-the-Middle Attack Detection in Advanced Metering Infrastructure

Smart grid (SG) engages bidirectional energy and data flow with advanced metering infrastructure (AMI) and lures attackers to exploit vulnerabilities in the critical infrastructure. Of the existing cyber threats, one of the major attacks is a man-in-the-middle (MITM) attack which eavesdrops and interrupts the communication between smart meter (SM) and data concentrator (DC) thereby paving the way for other cyber threats. Though many research works dealt with the solution to this attack, the majority of them are either simulation based or not intended for AMI. Hence, the proposal aims to develop a robust machine learning (ML) based intrusion detection system (IDS) to detect the attack in the real-time infrastructure. To attain this, AMI is setup using lamp loads, resistive loads, sensors, AtMega controller, Raspberry Pi, and a Server using the communication protocols namely Modbus TCP/IP and message queuing telemetry transport (MQTT). Then, the attack is emulated using Ettercap in a real-time testbed. The packets are captured with Wireshark and processed using preprocessing methods in Python. Furthermore, a pool of ML algorithms is extensively trained and validated on the practical testbed to identify suitable intrusion detection models. The results exhibit that the random forest (RF) algorithm outperforms the rest of the algorithms. Furthermore, to optimize its performance, the conventional RF algorithm is blended with a modified hybrid bat algorithm (HBA). The performance metrics of the hybrid model exhibit its supremacy over its counterparts. Proposed model video link:
https://drive.google.com/file/d/1Yq6nuNlr98klaoy07k83m4ega
SHOJIM0/view?usp = drive_link.