2021/04/06 by A. Sadik Satir, Umut Demir, Satir, A. Sadik +5
Engineering · #93-10 #Artificial Intelligence (cs.AI) #Computational Fluid Dynamics and Aerodynamics #FOS: Computer and information sciences #FOS: Electrical engineering #FOS: Mathematics #Guidance and Control Systems #I.2.6 #Military Defense Systems Analysis #Optimization and Control (math.OC) #Systems and Control (eess.SY) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2104.02491
openalex publication_date 2021/04/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
In this work, we propose a novel missile guidance algorithm that combines\ndeep learning based trajectory prediction with nonlinear model predictive\ncontrol. Although missile guidance and threat interception is a well-studied\nproblem, existing algorithms' performance degrades significantly when the\ntarget is pulling high acceleration attack maneuvers while rapidly changing its\ndirection. We argue that since most threats execute similar attack maneuvers,\nthese nonlinear trajectory patterns can be processed with modern machine\nlearning methods to build high accuracy trajectory prediction algorithms. We\ntrain a long short-term memory network (LSTM) based on a class of simulated\nstructured agile attack patterns, then combine this predictor with quadratic\nprogramming based nonlinear model predictive control (NMPC). Our method, named\nnonlinear model based predictive control with target acceleration predictions\n(NMPC-TAP), significantly outperforms compared approaches in terms of miss\ndistance, for the scenarios where the target/threat is executing agile\nmaneuvers.\n