Matrix completion, which was proposed by Candes and Tao, is a new matrix recovery technqiue which is come from compressed sensing (CS) theory. In general, Matrix reconstruction from few entries is impossiple. However, there exist recovery possiblity, if we know the Matrix has low-rank perpety. As Candes said, matrix completion will become a hot topic which can be employed many areas such as GPS and MIMO channel estimation.
MIMO channel estimation is a challenge task, especially in frequency-selective fading channel. Thus, MIMO channel estimation employ matrix completion which based on CS theory is a new idea. But, how we can do it? This is my research at present. If you have some suggestion about it, don't forget tell me. Waiting for your commances. Many thanks.