Ionospheric gradient anomaly detection using kernel support vector machine for ground-based augmentation systems

The ground-based augmentation system (GBAS) augments global navigation satellite systems (GNSS) to help high-level precision models as well as aircraft landings. The GBAS's performance is affected by gradients in ionospheric between the aircraft and reference stations. The ionospheric anomaly (IA) is an asymmetrical variation of the ionosphere that results in latent intimidations for GBAS. The most crucial task of automatic landing of aircraft speedily detects IA from the background noise, so as to intersect the security necessities. The aim of significant task is to quickly detect the IA from environmental noise to encounter the safety-critical necessities for GBAS. A kernel support vector machine (KSVM) approach was proposed for efficient detection of ionospheric anomalies and to improve detection speed. An offline-online method based on KSVM is suggested for extracting relevant data associated with anomaly features in the presence of noise. The proposed KSVM-based method demonstrates improved performance with a standard deviation of 0.8 at a change rate of 0.01 m/s, and achieves better detection times across various K-fold validation sets, with the best results observed at K=5. The findings suggest that the proposed approach enhances GBAS integrity, while simulations show a 50% improvement in average detection speed. Additionally, the KSVM approach is more sensitive and adaptable, offering superior detection rates and interpretability compared to current state-of-the-art techniques.