Statistical Detection Theory Statistical Detection Theory Il Wenhui Xiong NCL UESTC whxiong@uestc.edu.cn
whxiong@uestc.edu.cn Statistical Detection Theory Statistical Detection Theory II Wenhui Xiong NCL UESTC
System Model ●Summary of NP rule: ·Hypothesis test Deterministic->known signal Random signal->known pdf How about the hypothesis test with unknown or partially known pdf? Composite test whxiong@uestc.edu.cn 2
whxiong@uestc.edu.cn System Model 2 Summary of NP rule: Hypothesis test Deterministic -> known signal Random signal ->known pdf How about the hypothesis test with unknown or partially known pdf? Composite test
Composite Test-One-sided Test E.g.,Positive DC in WGN with known noise statistics Ho xIn]wln] A>0 H xIn]A+wln] n=0,1..W-1 whxiong@uestc.edu.cn
whxiong@uestc.edu.cn Composite Test—One-sided Test H1 x[n] = A + w[n] E.g., Positive DC in WGN with known noise statistics H0 x[n] = w[n] n=0,1…N-1 A>0
Composite Test-One-sided Test E.g.,Positive DC in WGN with known noise statistics Ho xIn]wln] A>0 H xIn]=A+wln] n=01..W-1 Apply NP rule P(X;A,H1) exp(-∑(xnl-A2 P(x:A,Ho) >Y exp(-2a∑2[n]) whxiong@uestc.edu.cn 4
whxiong@uestc.edu.cn Composite Test—One-sided Test 4 H1 x[n] = A + w[n] E.g., Positive DC in WGN with known noise statistics H0 x[n] = w[n] n=0,1…N-1 Apply NP rule A>0 P(x; A; H1) P(x; A; H0) = exp (¡ 1 2¾2 P(x[n] ¡ A) 2 exp(¡ 1 2¾2 Px 2[n]) > °
Composite Test-One-sided Test E.g.,Positive DC in WGN with known noise statistics Ho xIn]wln] A>0 H xIn]=A+wln] n=01..W-1 Apply NP rule P(x;A,H1) exp(-2石∑(cn-A2 P(x;A,Ho) >Y exp(-2a∑2[m]) -京(-2A∑n+N4)>1n whxiong@uestc.edu.cn
whxiong@uestc.edu.cn Composite Test—One-sided Test H1 x[n] = A + w[n] E.g., Positive DC in WGN with known noise statistics H0 x[n] = w[n] n=0,1…N-1 Apply NP rule A>0 P(x; A; H1) P(x; A; H0) = exp (¡ 1 2¾2 P(x[n] ¡ A) 2 exp(¡ 1 2¾2 Px 2[n]) > ° ¡ 1 2¾2 ³ ¡2A Xx[n] + NA 2 ´ > ln°