A dendritic cell mechanism for detection, identification, and evaluation of aircraft failures

Successful fault-tolerant control strategies often require vital tools that can accurately detect the failure, identify its root cause, and evaluate its nature and severity. Most of the existing methodologies in the field of failure detection, identification, and evaluation are limited to few subsystems with reduced number of features. Due to the complexity and multidimensionality of the aircraft system, new methodologies that are robust, accurate, and fast enough need to be developed for such systems. The biological immune system is a natural system that possesses vigorous peculiarities in protecting the mammalian body from harmful intruders and, therefore, may represent a rich source of inspiration to solve anomaly problems. This paper presents a novel integrated scheme for aircraft sub-system failure detection, identification, and evaluation based on the functionality of the biological dendritic cells and their interactions with the various components of the immune system. The proposed approach relies on using the self/nonself discrimination principle with the hierarchical multiself strategy to overcome the multidimensionality issues. The information collected by the artificial dendritic cells is fused in a way that convert the identification and evaluation problem into a pattern recognition problem. The proposed scheme was successfully tested for a supersonic fighter aircraft in a motion-based flight simulator with high detection, identification, and evaluation rates and practically zero false alarms.