Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I have the decision problem for 4 hypotheses as follows: $$H_j: Y_k=N_k-s_{jk},\ k=1,2,\ldots,n;\ j=0,1,2,3.$$ where signals are $s_{jk}=E_0\sin(w_cT(k-1)+(j+\frac{1}{2})\frac{\pi}{2}).$ $$$$ In vector form: $$\equiv H_{j}: \underline{Y}=\underline{N}+\underline{s}_j;\ j=0,1,2,3.$$ $$$$ How can I find the minimum error probability for equally likely signals in i.i.d. $N(0,\sigma^2)$ noise. (Thess signals are not orthogonal). how can I cobtain orthonormal signals for solving this problem? Thank you in advance.

share|improve this question
1  
Since you are a new user, here are some tips to get you started: Show your basic thoughts on the subject or things you have tried, or parts where you are confused. If an acceptable answer is posted for you, don't forget to accept the answer by clicking the check-mark next to their post so that way credit is properly assigned. –  MSEoris Nov 23 '12 at 16:37
add comment

1 Answer

Conditioned on the $j$-th signal being transmitted, the likelihood function of the observation $(Y_1, \ldots, Y_n)$ is proportional to $$\exp\left(-\frac{1}{2}\sum_{k=1}^n (s_{kj}-Y_k)^2\right)$$ Since the signals are equally likely to be transmitted, the_ minimum-error-probability decision rul_e is the same as the maximum-likelihood decision rule, viz.

Choose the hypothesis that has the largest likelihood

which in this instance means deciding that the signal $s_j$ that is closest in Euclidean metric to the observation $Y$ is the one that is most likely to have been transmitted. In other words, compute the four sums $$Z_j = \sum_{k=1}^n (s_{kj}-Y_k)^2, ~~j = 0, 1, 2, 3$$ and decide that signal $s_j$ was transmitted if $Z_j < \min_{i: i\neq j} Z_i$

share|improve this answer
    
@ Dilip Sarwate: Thank you for your answer! However, my main problem is in the calculation minimum error probability($P_e$) where $P_c=1-P_e=P_0(Z_0>Z_1,Z_0>Z_2,Z_0>Z_3)$. Because the signals ($s_j$) are not ortogonal to have independecies between $Z_j$. –  lewer Nov 23 '12 at 19:21
    
@lewer You need to state very carefully what you mean by minimum error probability. The minimum average error probability? The conditional minimum error probability given hat a specific $s_i$ was transmitted? The smallest probability that a given transmitted $s_i$ was demodulated as specific $s_j$? As a simple answer, read about the union bound or Bonfderroni inequalities. –  Dilip Sarwate Nov 23 '12 at 23:09
    
@lewer See my lecture notes for some ideas on how to approach the problem. –  Dilip Sarwate Nov 23 '12 at 23:18
    
@ Prof.Sarwate: Thank you for providing me your lecture notes in this field. My goal of minimum error probability is The conditional minimum error probability given that a specific $s_j$ was transmitted(on page.162 of your lecture notes)? –  lewer Nov 24 '12 at 6:22
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.