《超低頻非高斯參數估計和套用》是2014年6月國防工業出版社出版的圖書，作者蔣宇中。
基本介紹
- 中文名：超低頻非高斯參數估計和套用
- 作者：蔣宇中
- 出版時間：2014年6月
- 出版社：國防工業出版社
- 頁數：150 頁
- ISBN：9787118096521
- 定價：86.00
- 開本：16 開
- 裝幀：平裝
- 版次：1版1次
- 字數：208
- 中圖分類：TP84
內容簡介,目錄,
內容簡介
《超低頻非高斯噪聲模型及套用》介紹和討論了低頻（超低頻）無線通信非高斯噪聲建模和參數估計原理與方法，系統地介紹非高斯噪聲建模的數學物理基礎，參數估計的方法和主要關鍵技術。內容有：非高斯噪聲的經驗模型和統計物理模型，低頻大氣噪聲幅度機率分布模式的辨識，非高斯噪聲環境中的*接收機，噪聲模型的參數估計三種方法：特徵函式譜法、馬爾可夫鏈蒙特卡羅法和重要性採樣。最後介紹超低頻實驗接收機工程實現的若干重要問題。
本書可供從事軍事通信、低頻無線電研究、聲納與雷達信號處理等領域的研究人員和工程技術人員參考，也可作為高等學校信號處理和相關專業的本科生和研究生的參考用書。
目錄
{第 1 章}緒論}{{rm 1}}
{1.1}超低頻非高斯噪聲模型研究的意義}{1}
{1.2}超低頻大氣噪聲}{3}
{1.3}非高斯噪聲模型簡要歷史回顧}{6}
{1.3.1}經驗噪聲模型}{6}
{1.3.2}統計物理噪聲模型}{8}
{1.4}噪聲模型的參數估計}{10}
{第 2 章}非高斯噪聲理論模型的建立}{{rm 11}}
{2.1}Class A 和 Class B 噪聲模型}{11}
{2.2}Class A 和 Class B 噪聲模型的數學表達式}{12}
{2.2.1}廣義基本噪聲模型}{12}
{2.2.2}Class A 和 Class B 幅度機率分布}{13}
{2.3}$alpha $ 穩定過程}{15}
{2.3.1}$alpha $ 穩定分布定義}{15}
{2.3.2}$alpha $ 穩定分布特例}{16}
{2.3.3}$alpha $ 穩定分布矩性質}{17}
{第 3 章}超低頻大氣噪聲幅度機率分布模式的辨識}{{rm 19}}
{3.1}引言}{19}
{3.2}超低頻大氣噪聲建模的理論依據}{19}
{3.3}超低頻大氣噪聲測量設備簡介}{20}
{3.4}超低頻信道電磁噪聲數據的測量分析方法}{22}
{3.5}超低頻信道電磁噪聲數據的預處理}{23}
{3.6}超低頻信道電磁噪聲數據的非常態分配檢驗}{25}
{3.7}寬頻超低頻信道大氣噪聲數據的幅度統計特性分析}{25}
{3.8}窄帶超低頻信道大氣噪聲數據的幅度統計特性分析}{29}
{3.9}結論}{33}
{第 4 章}非高斯噪聲中的信號檢測}{{rm 34}}
{4.1}最優接收機結構}{34}
{4.1.1}局部最佳接收機結構}{36}
{4.1.2}次優接收機結構}{37}
{4.1.3}非線性處理器的性能估計}{38}
{4.2}非高斯噪聲環境下 Turbo 碼的解碼算法性能分析及改進}{39}
{4.2.1}Class A 噪聲模型的簡化}{39}
{4.2.2}非高斯噪聲 LOG-MAP-CA 算法推導}{41}
{4.2.3}仿真結果及討論}{43}
{4.3}噪聲的非參數機率密度估計}{45}
{4.4}機率密度與非線性函式聯合估計方法}{49}
{4.5}本章 MATLAB 仿真程式及子程式}{52}
{第 5 章}基於特徵函式譜噪聲模型的參數估計}{{rm 57}}
{5.1}引言}{57}
{5.2}基於特徵函式的 Class A 參數估計算法}{58}
{5.2.1}Class A 模型的特徵函式}{58}
{5.2.2}基於特徵函式的 Class A 模型參數估計算法的推導}{59}
{5.2.3}基於特徵函式的 Class A 模型參數估計算法仿真}{60}
{5.3}基於特徵函式求逆的最大似然參數估計算法}{61}
{5.3.1}機率密度函式的 FFT 形式}{62}
{5.3.2}最大似然估計}{64}
{5.3.3}實驗仿真}{65}
{5.4}Class B 噪聲模型的參數估計}{66}
{5.4.1}Class B 噪聲模型}{66}
{5.4.2} Class B 噪聲模型參數估計算法}{67}
{5.4.3}Class B 參數估計算法的仿真及結論}{69}
{5.5}Class B 噪聲模型的非線性回歸估計}{71}
{5.5.1}非線性回歸估計}{72}
{5.5.2}初始值估計和 ${lambda _k}$ 序列的生成}{73}
{5.5.3}仿真及實際結果}{74}
{5.6}本章 MATLAB 仿真程式及子程式}{76}
{第 6 章}馬爾可夫鏈蒙特卡羅法噪聲模型的參數估計}{{rm 82}}
{6.1}混合模型的 MCMC 參數估計}{82}
{6.1.1}$alpha $ 穩定分布乘法性質}{82}
{6.1.2}混合模型}{83}
{6.1.3}貝葉斯層次模型和先驗}{83}
{6.1.4}MCMC 算法}{84}
{6.1.4.1}通過 Gibbs 抽樣更新權重係數 ${bm {w}}$}{85}
{6.1.4.2}通過 Gibbs 抽樣更新參數 $gamma $}{85}
{6.1.4.3}通過 Gibbs 抽樣更新參數 $sigma ^{-2}$}{85}
{6.1.4.4}通過 Metropolis-Hastings 算法更新 $alpha $}{85}
{6.1.4.5}通過 Gibbs 抽樣更新參數 $kappa ,beta $}{86}
{6.1.4.6}更新標籤變數 $z$}{86}
{6.1.4.7}通過 Metropolis-Hastings 算法更新變數 $lambda $}{86}
{6.1.5}仿真及實測結果}{87}
{6.2}基於最大後驗機率的 Class A 參數估計算法}{87}
{6.2.1}基於最大後驗機率參數估計算法推導}{88}
{6.2.2}基於最大後驗機率的 Class A 參數估計算法仿真}{90}
{6.3}本章 MATLAB 仿真程式及子程式}{92}
{第 7 章}兩維的 M-Class A 噪聲模型的參數估計}{{rm 97}}
{7.1}引言}{97}
{7.2}多維 M-Class A 噪聲模型}{98}
{7.3}M-Class A 噪聲模型參數的馬爾可夫鏈蒙特卡羅估計算法}{101}
{7.3.1}M-Class A 噪聲模型參數的貝葉斯估計算法推導}{101}
{7.3.2}M-Class A 模型參數估計算法的計算機仿真及結論}{105}
{7.4}M-Class A 噪聲模型的 PMC 參數估計算法}{106}
{7.4.1}PMC 參數估計算法的推導}{107}
{7.4.2}PMC 算法的 CUDA 並行計算}{109}
{7.5}本章 MATLAB 仿真程式及子程式}{115}
{第 8 章}超低頻實驗接收機實現的若干重要問題}{{rm 123}}
{8.1}引言}{123}
{8.2}編解碼方案設計}{123}
{8.2.1}編解碼體系的選擇}{124}
{8.2.2}編碼器的結構參數}{125}
{8.2.3}編碼器的生成多項式}{125}
{8.2.4}解碼算法}{127}
{8.3}調製解調方案設計}{127}
{8.4}抗干擾設計}{129}
{8.5}同步與交織}{130}
{8.6}接收機信號處理設計方案}{133}
{8.7}結語}{135}
{附錄 Ahskip 1emrelax Class A 和 Class B 模型的幅度機率分布推導}{{rm 136}}
{A.1hskip 1emrelax 窄帶接收機情形}{137}
{A.2hskip 1emrelax Class A 特徵函式}{139}
{A.3hskip 1emrelax Class B 特徵函式}{140}
{A.4hskip 1emrelax Class A 和 Class B 模型的幅度機率分布}{142}
{參考文獻}{{rm 144}}
contentsfinish
chapter*{fzxbs Contents}
markboth{Contents}{Contents}
thispagestyle{empty}
{1}Introduction}{{rm 1}}
{1.1}The Need of Non-Gaussian noise models}{1}
{1.2}Atmospheric Noise in Extremely Low Frequency}{3}
{1.3}Brief Review of Non-Gaussian noise models}{6}
{1.3.1}Empirical noise models}{6}
{1.3.2}Statistical and physical noise models}{8}
{1.4}Parameters estimation of noise models}{10}
{2}Development of Non-Gaussian Noise Models}{{rm 11}}
{2.1}Class A and Class B noise models}{11}
{2.2}Mathematical Expression of Class A and Class B noise models}{12}
{2.2.1}Basic Middleton noise model}{12}
{2.2.2}Amplitude Probability Distribution of Class A and Class B models}{13}
{2.3}$alpha $ stable process}{15}
{2.3.1}The Definition of $alpha$ stable process}{15}
{2.3.2}Special cases of $alpha $ stable distributions}{16}
{2.3.3}Moments of $alpha$ stable distributions}{17}
{3}The Analysis of the Amplitude Probabilityhspace*{7.3mm}Distribution of ELF Atmospheric Noises}{{rm 19}}
{3.1}Introduction}{19}
{3.2}The Theory of the development of ELF atmospheric noises}{19}
{3.3}The Synopsis of ELF atmospheric noise measurement devices}{20}
{3.4}Measurement analysis methods for Electromagnetic Noise data inhspace*{8mm}ELF channels}{22}
{3.5}Pre-processor for Electromagnetic noise data}{23}
{3.6}The Non-normality Test for Electromagnetic noise data}{25}
{3.7}The Analysis of Statistical Characteristics of the Amplitude ofhspace*{8.2mm}the wideband ELF Atmospheric
noise data}{25}
{3.8}The Analysis of Statistical Characteristics of the Amplitude of thehspace*{8.2mm}narrowband ELF Atmospheric
noise data}{29}
{3.9}Conclusion}{33}
{4}Signal Detection in Non-Gaussian noise}{{rm 34}}
{4.1}The Scheme of the optimum receiver}{34}
{4.1.1}The Scheme of the locally-optimum receiver}{36}
{4.1.2}The Scheme of the sub-optimum receiver}{37}
{4.1.3}Performances of the non-linear processors}{38}
{4.2}Performances Improvement of Turbo code in Non-Gaussian noise}{39}
{4.2.1}The Simplification of the Class A noise model}{39}
{4.2.2}The Derivative of the LOG-MAP-CA algorithm in Non-Gaussianhspace*{9.8mm}noise models}{41}
{4.2.3}Simulation Results and Analysis}{43}
{4.3}Non-parameters Estimation of the Probability Density Functionhspace*{8.2mm}of noise models}{45}
{4.4}Joint Estimation of the Probability Density Function and thehspace*{8.2mm}Non-linear Function}{49}
{4.5}Related Matlab code}{52}
{5}Parameters Estimation of noise models based onhspace*{7.3mm}Characteristic Functions}{{rm 57}}
{5.1}Introduction}{57}
{5.2}Parameters Estimation of the Class A model based on thehspace*{8mm}Characteristic Function}{58}
{5.2.1}The Characteristic Function of the Class A model}{58}
{5.2.2}Development of the Algorithm}{59}
{5.2.3}Simulation Results and Analysis}{60}
{5.3}Maximum Likelihood Estimation of the Class A model's parametershspace*{8.2mm}based on the Characteristic
Function}{61}
{5.3.1}FFT Form of the Probability Density Function}{62}
{5.3.2}Maximum Likelihood Estimation}{64}
{5.3.3}Simulation Results and Analysis}{65}
{5.4}Parameters Estimation of the Class B model}{66}
{5.4.1}Class B noise model}{66}
{5.4.2}Development of the algorithm}{67}
{5.4.3}Simulation Results and Analysis}{69}
{5.5}Nonlinear Regression-Type Estimation of the Parameters of thehspace*{8mm}Class B Noise Model}{71}
{5.5.1}Nonlinear Regression-Type Estimation}{72}
{5.5.2}The Estimation of the initial value and the Generation of ${lambda _k}$}{73}
{5.5.3}Simulation Results and Analysis}{74}
{5.6}本章 MATLAB 仿真程式及子程式}{76}
{第 6 章}馬爾可夫鏈蒙特卡羅法噪聲模型的參數估計}{{rm 82}}
{6.1}混合模型的 MCMC 參數估計}{82}
{6.1.1}$alpha $ 穩定分布乘法性質}{82}
{6.1.2}混合模型}{83}
{6.1.3}貝葉斯層次模型和先驗}{83}
{6.1.4}MCMC 算法}{84}
{6.1.4.1}通過 Gibbs 抽樣更新權重係數 ${bm {w}}$}{85}
{6.1.4.2}通過 Gibbs 抽樣更新參數 $gamma $}{85}
{6.1.4.3}通過 Gibbs 抽樣更新參數 $sigma ^{-2}$}{85}
{6.1.4.4}通過 Metropolis-Hastings 算法更新 $alpha $}{85}
{6.1.4.5}通過 Gibbs 抽樣更新參數 $kappa ,beta $}{86}
{6.1.4.6}更新標籤變數 $z$}{86}
{6.1.4.7}通過 Metropolis-Hastings 算法更新變數 $lambda $}{86}
{6.1.5}仿真及實測結果}{87}
{6.2}基於最大後驗機率的 Class A 參數估計算法}{87}
{6.2.1}基於最大後驗機率參數估計算法推導}{88}
{6.2.2}基於最大後驗機率的 Class A 參數估計算法仿真}{90}
{6.3}本章 MATLAB 仿真程式及子程式}{92}
{第 7 章}兩維的 M-Class A 噪聲模型的參數估計}{{rm 97}}
{7.1}引言}{97}
{7.2}多維 M-Class A 噪聲模型}{98}
{7.3}M-Class A 噪聲模型參數的馬爾可夫鏈蒙特卡羅估計算法}{101}
{7.3.1}M-Class A 噪聲模型參數的貝葉斯估計算法推導}{101}
{7.3.2}M-Class A 模型參數估計算法的計算機仿真及結論}{105}
{7.4}M-Class A 噪聲模型的 PMC 參數估計算法}{106}
{7.4.1}PMC 參數估計算法的推導}{107}
{7.4.2}PMC 算法的 CUDA 並行計算}{109}
{7.5}本章 MATLAB 仿真程式及子程式}{115}
{第 8 章}超低頻實驗接收機實現的若干重要問題}{{rm 123}}
{8.1}引言}{123}
{8.2}編解碼方案設計}{123}
{8.2.1}編解碼體系的選擇}{124}
{8.2.2}編碼器的結構參數}{125}
{8.2.3}編碼器的生成多項式}{125}
{8.2.4}解碼算法}{127}
{8.3}調製解調方案設計}{127}
{8.4}抗干擾設計}{129}
{8.5}同步與交織}{130}
{8.6}接收機信號處理設計方案}{133}
{8.7}結語}{135}
{附錄 Ahskip 1emrelax Class A 和 Class B 模型的幅度機率分布推導}{{rm 136}}
{A.1hskip 1emrelax 窄帶接收機情形}{137}
{A.2hskip 1emrelax Class A 特徵函式}{139}
{A.3hskip 1emrelax Class B 特徵函式}{140}
{A.4hskip 1emrelax Class A 和 Class B 模型的幅度機率分布}{142}
{參考文獻}{{rm 144}}
contentsfinish
chapter*{fzxbs Contents}
markboth{Contents}{Contents}
thispagestyle{empty}
{1}Introduction}{{rm 1}}
{1.1}The Need of Non-Gaussian noise models}{1}
{1.2}Atmospheric Noise in Extremely Low Frequency}{3}
{1.3}Brief Review of Non-Gaussian noise models}{6}
{1.3.1}Empirical noise models}{6}
{1.3.2}Statistical and physical noise models}{8}
{1.4}Parameters estimation of noise models}{10}
{2}Development of Non-Gaussian Noise Models}{{rm 11}}
{2.1}Class A and Class B noise models}{11}
{2.2}Mathematical Expression of Class A and Class B noise models}{12}
{2.2.1}Basic Middleton noise model}{12}
{2.2.2}Amplitude Probability Distribution of Class A and Class B models}{13}
{2.3}$alpha $ stable process}{15}
{2.3.1}The Definition of $alpha$ stable process}{15}
{2.3.2}Special cases of $alpha $ stable distributions}{16}
{2.3.3}Moments of $alpha$ stable distributions}{17}
{3}The Analysis of the Amplitude Probabilityhspace*{7.3mm}Distribution of ELF Atmospheric Noises}{{rm 19}}
{3.1}Introduction}{19}
{3.2}The Theory of the development of ELF atmospheric noises}{19}
{3.3}The Synopsis of ELF atmospheric noise measurement devices}{20}
{3.4}Measurement analysis methods for Electromagnetic Noise data inhspace*{8mm}ELF channels}{22}
{3.5}Pre-processor for Electromagnetic noise data}{23}
{3.6}The Non-normality Test for Electromagnetic noise data}{25}
{3.7}The Analysis of Statistical Characteristics of the Amplitude ofhspace*{8.2mm}the wideband ELF Atmospheric
noise data}{25}
{3.8}The Analysis of Statistical Characteristics of the Amplitude of thehspace*{8.2mm}narrowband ELF Atmospheric
noise data}{29}
{3.9}Conclusion}{33}
{4}Signal Detection in Non-Gaussian noise}{{rm 34}}
{4.1}The Scheme of the optimum receiver}{34}
{4.1.1}The Scheme of the locally-optimum receiver}{36}
{4.1.2}The Scheme of the sub-optimum receiver}{37}
{4.1.3}Performances of the non-linear processors}{38}
{4.2}Performances Improvement of Turbo code in Non-Gaussian noise}{39}
{4.2.1}The Simplification of the Class A noise model}{39}
{4.2.2}The Derivative of the LOG-MAP-CA algorithm in Non-Gaussianhspace*{9.8mm}noise models}{41}
{4.2.3}Simulation Results and Analysis}{43}
{4.3}Non-parameters Estimation of the Probability Density Functionhspace*{8.2mm}of noise models}{45}
{4.4}Joint Estimation of the Probability Density Function and thehspace*{8.2mm}Non-linear Function}{49}
{4.5}Related Matlab code}{52}
{5}Parameters Estimation of noise models based onhspace*{7.3mm}Characteristic Functions}{{rm 57}}
{5.1}Introduction}{57}
{5.2}Parameters Estimation of the Class A model based on thehspace*{8mm}Characteristic Function}{58}
{5.2.1}The Characteristic Function of the Class A model}{58}
{5.2.2}Development of the Algorithm}{59}
{5.2.3}Simulation Results and Analysis}{60}
{5.3}Maximum Likelihood Estimation of the Class A model's parametershspace*{8.2mm}based on the Characteristic
Function}{61}
{5.3.1}FFT Form of the Probability Density Function}{62}
{5.3.2}Maximum Likelihood Estimation}{64}
{5.3.3}Simulation Results and Analysis}{65}
{5.4}Parameters Estimation of the Class B model}{66}
{5.4.1}Class B noise model}{66}
{5.4.2}Development of the algorithm}{67}
{5.4.3}Simulation Results and Analysis}{69}
{5.5}Nonlinear Regression-Type Estimation of the Parameters of thehspace*{8mm}Class B Noise Model}{71}
{5.5.1}Nonlinear Regression-Type Estimation}{72}
{5.5.2}The Estimation of the initial value and the Generation of ${lambda _k}$}{73}
{5.5.3}Simulation Results and Analysis}{74}
{5.6}Related Matlab code}{76}
{6}Parameters Estimation of noise models throughhspace*{7.1mm}MCMC algorithms}{{rm 82}}
{6.1}Parameters Estimation of the Mixture models by MCMC algorithms}{82}
{6.1.1}The Product Property of the $alpha $ stable distribution}{82}
{6.1.2}The Mixture Model}{83}
{6.1.3}Bayesian Hierarchical Model and Prior Distributions}{83}
{6.1.4}MCMC algorithms}{84}
{6.1.4.1}Updating weights ${bm {w}}$ through the Gibbs sampling}{85}
{6.1.4.2}Updating parameters $gamma$ through the Gibbssampling}{85}
{6.1.4.3}Updating parameters $sigma^{-2}$ through the Gibbssampling}{85}
{6.1.4.4}Updating parameters $alpha$ through the Metropolis-Hastings sampling}{85}
{6.1.4.5}Updating parameters $kappa ,beta $ through the Gibbssampling}{86}
{6.1.4.6}Updating the label variables $z$}{86}
{6.1.4.7}Updating the variables $lambda$ through the Metropolis-Hastings sampling}{86}
{6.1.5}Simulation Results and Analysis}{87}
{6.2}MAP Estimation of Parameters of the Class A model}{87}
{6.2.1}Development of the algorithm}{88}
{6.2.2}Simulation Results and Analysis}{90}
{6.3}Related Matlab code}{92}
{7}Parameters Estimation of the M-Class A noise model}{{rm 97}}
{7.1}Introduction}{97}
{7.2}M-Class A noise model}{98}
{7.3}Parameters Estimation of the M-Class A noise model byhspace*{8mm}MCMC algorithms}{101}
{7.3.1}Development of the algorithm}{101}
{7.3.2}Simulation Results and Analysis}{105}
{7.4}Parameters Estimation of the M-Class A noise model by PMChspace*{8.2mm}algorithms}{106}
{7.4.1}Population-based Monte Carlo algorithm}{107}
{7.4.2}Parallel Computation Design of the algorithm by CUDA schemes}{109}
{7.5}Related Matlab code}{115}
{8}Discussion about some topics on the design ofhspace*{7.3mm}ELF receivers}{{rm 123}}
{8.1}Introduction}{123}
{8.2}The Design of Encoders and Decoders}{123}
{8.2.1}The Choice of Encoders}{124}
{8.2.2}The Scheme of Encoders}{125}
{8.2.3}The Generator Polynomials of Encoders}{125}
{8.2.4}The Algorithm of Decoders}{127}
{8.3}The Design of Modulators and Demodulators}{127}
{8.4}The Anti-interference Design}{129}
{8.5}Synchronizer and Interleaver}{130}
{8.6}Signal processing in receivers}{133}
{8.7}Conclusion}{135}
{Appendix Ahskip 1emrelax The derivative of the amplitude probabilityhspace*{8.2mm}distribution of Class A and
Class B model}{{rm 136}}
{A.1hskip 1emrelax The narrow receiver conditions}{137}
{A.2hskip 1emrelax The characteristic function of the Class A model}{139}
{A.3hskip 1emrelax The characteristic function of the Class B model}{140}
{A.4hskip 1emrelax The amplitude probability distribution of Class A and Class B model}{142}
"
