Complex infomax for blind separation of delayed sources

The information maximization (Infomax) based on information entropy theory is a class of methods that can be used to blindly separate mixed signals. In independent component analysis (ICA), the Infomax criterion is proposed by Bell and Sejnowski. Torkkola applied this criterion to separate the delayed sources blindly, which is more approximate to the real world than ICA model. However, Torkkola´s method works only in the real-valued field. In this paper, we extend the Infomax for blind separation of delayed sources to the complex case and derive the adaptation rules based on gradient ascent. The simulation results illustrate the performance of the studied algorithm.