Bio-Inspired Intelligent Anomaly Diagnosis System for Health Monitoring in Connected Vehicles

The security of Connected Autonomous Vehicles (CAVs) is critical due to their susceptibility to anomalies, including physical faults and cyberattacks. Available solutions to tackle such anomalies are often limited in detection capability or computational efficiency. This paper introduces a novel application of the Negative Selection Algorithm (NSA) for anomaly detection in CAVs, inspired by the human immune system’s ability to distinguish self from nonself. In addition, critical NSA parameters are analyzed and tuned for optimal performance in this new application. Our framework considers both internal and external anomalies in a platoon of vehicles and is evaluated using various scenarios and a physical testbed including QCars. The obtained experimental results indicate the effectiveness and reliability of the designed framework for identifying abnormal behaviors in CAVs while addressing the limitations of traditional methods.