基于深度Q學(xué)習(xí)的組網(wǎng)雷達(dá)閃爍探測(cè)調(diào)度方法

打開(kāi)文本圖片集

中圖分類(lèi)號(hào)： TN973 文獻(xiàn)標(biāo)志碼：A DOI：10.12305/j.issn.1001-506X.2025.05.07

Abstract：The neted radar scintilation detection system can improve the cooperative detection performance and survival rate of radar.It is an urgent problem to select a suitableradar cooperative detection startup and limit the startup exposure time of a single radar to adapt to the ever-changing environmental threats. In this regard，a netted radar scintilltion detection scheduling method is presented based on deep Q-learning（DQL） reinforcement learning algorithmto limit the startup time of a single radar.Firstly，the threat degree model of theair jammer tothe netted radar and the scintillation detection modelof the netted radar to the air jammer are established.Then，thereinforcement learning reward functionof the threat degree and the neted scintilation detection probability is proposed.Finally，the optimal scintilation startup decision scheduling scheme of the netted radar is obtained by using the proposed DQL algorithm.The simulation results show that the average benefit rate of the proposed DQL scheduling method is superior to random scheduling，artificial bee colony scheduling and double deep Q network（DDQN） scheduling methods，and the scheduling response time is less.

Keywords：netted radar；scintillation detection；reinforcement learning；deep Q -learning（DQL）；double deep Q network（DDQN）

0 引言

空中干擾機(jī)執(zhí)行突防任務(wù)時(shí)搭載了機(jī)載電子干擾設(shè)備，在干擾突防過(guò)程中，空中干擾機(jī)會(huì)對(duì)地面雷達(dá)網(wǎng)實(shí)施電磁干擾，伺機(jī)收集地面雷達(dá)信息，單部雷達(dá)很難完成探測(cè)和抗干擾任務(wù)，也很難保證探測(cè)結(jié)果的可靠性和準(zhǔn)確性。（剩余13551字）