Open AccessApplication of differential countermeasure machine learning model in the field of UAV circuit inspection
Author(s) -
Xinrui Huang
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2083/3/032027
Subject(s) - trajectory , path (computing) , computer science , process (computing) , countermeasure , differential (mechanical device) , control theory (sociology) , artificial intelligence , field (mathematics) , point (geometry) , simulation , computer vision , mathematical optimization , mathematics , engineering , aerospace engineering , physics , geometry , control (management) , astronomy , programming language , operating system , pure mathematics
In this paper, we establish one-objective differential game equations for one-to-one attack and defense in a two-dimensional plane. Through calculation and visual analysis, obtain the optimal pursuit path movement trajectory, and through machine learning method to training UAV, using cycle process of simulation, output with time growth each cycle pursuit path results, by comparing the movement trajectory image of the pursuit results and find the sheep just escape critical point exit. After that, the angle difference between the two initial positions was changed and tested again to enable the UAV to learn the optimal escape strategy more comprehensively, thus making a more precise path selection. Finally, this method can be reasonably evaluated.