李燕+何怡剛+尹柏強
摘 要:針對復雜電磁環(huán)境定性分析和定量分級問題,提出基于廣義S變換與PNN(概率神經(jīng)網(wǎng)絡)的電磁環(huán)境復雜度評估方法.應用廣義S變換對電磁環(huán)境中受干擾的樣本信號進行時頻分析,同步提取時頻域時域占用度,頻域占用度和能量占用度評估參數(shù),分析了各參數(shù)的物理意義,并給出了各評估參數(shù)的計算公式.同時增加頻率重合度、調制格式相似度和背景噪聲強度等評估參數(shù),利用獲取的參數(shù)訓練PNN,并進行分類.仿真實驗證明基于廣義S變換與PNN的電磁環(huán)境復雜度評估方法在背景噪聲大、訓練樣本數(shù)少的條件下能有效提取評估參數(shù),并對電磁環(huán)境復雜度進行有效合理評估.
關鍵詞:電磁環(huán)境;復雜度;評估;S變換;PNN
中圖分類號: TM937 文獻標志碼:A
Electromagnetic Environment Complex Evaluation Based onS-transform and PNN
LI Yan1,HE Yigang1,2 ,YIN Baiqiang2
(1 .College of Electrical and Information Engineering ,Hunan University ,Changsha 410082,China;
2.School of Electrical Engineering and Automation,Hefei University of Technology,Hefei 230009,China;)
Abstract:For electromagnetic environment complex evaluation,a novel evaluation method was proposed based on S-transform and probabilistic neural network (PNN). S-transform was applied to perform time-frequency analysis on the electromagnetic environment disturbance samples,the evaluation parameters such as time occupation,frequency occupation and energy occupation of the samples were extracted synchronously,physical interpretation of each parameter was analyzed,and the calculation formulas of them were shown. At the same time,frequency coincidence,modulation similarity and background noise intensity were selected as important evaluation indicators. These parameters were then used to train a PNN which was adopted to evaluate environmental complexity. The simulation results show that the proposed method has integrity of parameter extraction and relatively high evaluation accuracy,and it can also give excellent evaluation for high level noises background and small training set.
Key words:electromagnetic environment;complexity;evaluation;S-transform;probability neural network
電磁環(huán)境復雜度是指在一定時間、一定空間內,電磁信號交集的復雜程度.電磁環(huán)境復雜度定性和定量分級是有效應對電磁雜波干擾、正確指揮決策、信息化戰(zhàn)爭的基礎.美國學者Tetley等人首次提出了基于數(shù)學模型的電磁環(huán)境分析預測方法,從數(shù)學分析角度給出了電磁環(huán)境的基本模型[1].目前,國內對電磁環(huán)境復雜度研究主要以“四域法”為基礎.文獻[2]提出用于描述戰(zhàn)場電磁環(huán)境“四域”特征指標,并給出具體指標計算方法.文獻[3]提出了一種基于層次分析方法(AHP)的電磁環(huán)境復雜度評估方法,細化了評估指標,明確了評估標準參數(shù),確定標度準則.文獻[4]提出一種基于D-S證據(jù)的評估方法,采用模糊數(shù)學法將定性的參數(shù)問題轉為定量描述.文獻[5]引入多元聯(lián)系數(shù),構建了電磁環(huán)境復雜度綜合評估方法,權值的設定根據(jù)均分原則.文獻[6]在四域法的基礎上增加了異動信號率作為評估指標,采用模糊推理系統(tǒng)對其進行評估.上述評估法存在以下不足:首先,特征參數(shù)時間占用度、頻率占用度和能量占用度作為三個獨立的參數(shù)分別進行計算,獨立的參數(shù)測量計算對系統(tǒng)的評估會造成較大的誤差[6].其次,采用專家系統(tǒng)及模糊規(guī)則對經(jīng)驗知識的依賴性強,具有一定的主觀性且自適應性較差[7].第三,忽略了評估對象,評估對象執(zhí)行效能的好壞才是評價電磁環(huán)境復雜度的最終指標[8].
針對這些問題,本文提出基于廣義S變換[9]和概率神經(jīng)網(wǎng)絡[10]的電磁環(huán)境評估方法.利用廣義S變換在時頻域內同步提取時域占用度、頻域占用度和能量占用度指標,同時,增加頻率重合度、調制方式和背景噪聲強度等對系統(tǒng)影響較大的評估指標,采用自適應PNN方法,對電磁環(huán)境復雜度進行評估.
1 廣義S變換域時頻參數(shù)的提取endprint