Optimized MSE estimation


Description:

Let X_1^n be a sequence of random variables with a known probability distribution p_X over R. Assume that we have a finite budget of taking m linear measurements from X_1^n and the idea is to estimate the value of the signal X_1^n after having the linear measurements. We are going to find the best possible set of linear measurements which give the minimal distortion possible. Also, we are interested to know how this optimal set depends on the distribution of the signal. Moreover, how is the stability of this optimal set to measurement noise? We are interested to study the problem in finite length and asymptotic regime and quantify the possible differences between the two cases.

Requirement:

A good background in Math, Good programming skill with Matlab or C/C++

Contact:

Saeid Haghighatshoar, saeid.haghighatshoar@epfl.ch

back to master semester projects menu

Last modified:: %2013/%01/%10 %09:%Jan