Consider a channel model y=hx+n,Where x is the QPSK symbol(i.e.either 1+i,1-i,-1-i,-1+i).Let n be an AWGN sample with zero mean and unit variance.Assume that we received a signal y=(0.5+0.5i), now to perform the Maximum Likelihood ML,these parameters are enough?

 

what system u r working on .... ur question is giving insufficient knowledge

 

It should be. The question can be looked at a couple of different ways. But what it comes down to is whether you can determine which of the four possible signals was most likely to have been sent given the actual signal that was received. What else do you think might be needed?

 

no I was thinking about system response h , but is it really required in determining the ML? that is the gist of problem.