Reliable anomaly detection in automatic flight control systems using a hybrid ABC–NSA approach

Purpose This study aims to develop a robust and efficient hybrid fault detection method for Automatic Flight Control Systems (AFCS) by integrating the Artificial Bee Colony (ABC) algorithm with the Negative Selection Algorithm (NSA). The goal is to improve anomaly detection performance while reducing computational cost in real-time flight data monitoring. Design/methodology/approach Control column position data from a Boeing 737 aircraft operated by a local airline are used as self-data in the NSA framework. The ABC algorithm is used to optimize the number and distribution of detectors using benchmark functions including Sphere, Elliptic and Sum Square, to enhance detector coverage and minimize redundancy. Three distance metrics – Euclidean, Manhattan and Infinity norm – are incorporated into the ABC–NSA algorithm to evaluate their impact on fault detection accuracy, false alarm rate, runtime and overall effectiveness. Findings Experimental results demonstrate that the Manhattan distance metric outperforms the others, achieving a fault detection rate of 99.18%, a false alarm rate of 1.11% and the highest Figure of Merit with an optimal detector count of 1,000. The ABC algorithm effectively reduces the detector population size without sacrificing accuracy, significantly decreasing the runtime compared to traditional NSA implementations. These improvements support reliable and timely detection of anomalies in flight control data. Originality/value This research presents the first application of ABC optimization for generating optimal NSA detector populations tailored for flight control anomaly detection. The hybrid ABC–NSA approach provides a novel, low-cost and computationally efficient solution suitable for integration into real-time flight control systems, thus contributing to enhanced flight safety and operational reliability.