Situational awareness has vital importance for next-generation fighter aircraft, and it enables pilots to assess, anticipate, and respond adeptly to dynamic combat scenarios. Pop-up threats may occur at any moment in the uncertainty of warfare. That's why, situational awareness calculations have to be executed in real-time, and the time complexity of situational awareness functions becomes critical. In this work, a machine learning-based clustering method that optimizes the time complexity of situational awareness functions with offline preprocessing of radar cross-section matrices is proposed. Survivability assessment is one of the crucial functionalities of situational awareness, and it plays a key role in modern air warfare. Radar cross-section is the most important low observability feature that is directly related to the survivability of aircraft because the probability of being detected or tracked affects survivability in an unfavorable manner. In order to assess survivability by taking radar cross-section into account through an air mission route, a significant number of iterations are required with raw radar cross-section matrices which represent the radar cross-section value for each pair of aspect azimuth and elevation angles. Clustering radar cross-section matrices into rectangle-like regions is a brilliant idea since stealth aircraft generally have sharp edges, flat surfaces, and rectangular shapes, and using clusters can eliminate the need for a large number of iterations. In this work, we proposed a machine learning-based clustering method for radar cross-section matrices; we showed how clustering radar cross-section matrices reduces the time complexity of survivability assessment calculations for next-generation fighter aircraft. Bias and mean bias error values for each cluster are calculated using the raw radar cross-section matrix as the ground truth and compared with other radar cross-section modeling approaches in the literature. Furthermore, performance benchmarks of survivability calculations through air mission routes are performed with raw radar cross-section matrices and clustered radar cross-section matrices.
Machine Learning Based Radar Cross-Section Clustering Towards Enhanced Situational Awareness for Next-Generation Fighter Aircraft
29.09.2024
2727935 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
The Next Step Towards Enhanced Situational Awareness
British Library Conference Proceedings | 1997
|The next step toward enhanced situational awareness
IEEE | 1997
|The jet fighter radar cross section
Tema Archiv | 1975
|The Jet Fighter Radar Cross Section
IEEE | 1975
|IMPROVING AIRCRAFT SITUATIONAL AWARENESS INSTRUMENTATION
Europäisches Patentamt | 2021