Font Size: a A A

Research On Anti-jamming Strategy Of Data Link Combat Platform Based On Reinforcement Learning

Posted on:2019-11-17Degree:MasterType:Thesis
Country:ChinaCandidate:C Q MaFull Text:PDF
GTID:2392330611993164Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
At present,data link equipment has developed a relatively complete system.However,there is still a big gap between the requirements of systematic combat with strong confrontation.Especially in terms of anti-jamming,as the confrontation becomes more and more fierce,existing anti-jamming measures and applications of data link can't fully meet the needs of combat conditions.Therefore,it is very important question to improve the anti-jamming ability in future data link system.In view of the above problems,this thesis takes the data link mobile combat platform as the research object,proposes to conduct research based on power control and route planning,mainly on route planning with reinforcement learning.This is an extension of the original power control method and the research content mainly includes:1.Modeling the problem as a discrete Markov decision process,designing a two-dimensional route planning environment model and a Bayesian-Stackelberg game data link power control strategy,and obtaining the Stackelberg equilibrium solution of the game.2.In the environment where the location of the interference source is fixed,designing a route planning algorithm suitable for the problem based on Deep Deterministic Policy Gradient algorithm,and reward shaping is added to solve the problem of reward sparseness.Finally,the combat platform adapts to the interference.3.In the environment where the interference source moves,applying meta reinforcement learning into route planning.Designing a dynamic adapting algorithm based on Model-Agnostic Meta-Learning algorithm and taking Policy Optimization with Penalized Point Probability Distance as policy optimizer.Finally realizing dynamic adaptive to the changing environment.4.The experimental results show that the proposed algorithm can improve the anti-jamming ability of the system to a certain extent,and can be used as an extension of the existing anti-jamming method,which has certain research significance for improving the electronic countermeasure of data link system.
Keywords/Search Tags:Data Link, Anti-jamming, Path Planning, Deep Reinforcement Learning, Meta Reinforcement Learning
PDF Full Text Request
Related items