Blind Signal Separation Based Co-channel Interference Suppression

In the wireless cellular communication system (such as the LTE system) whose frequency reuse factor is close to 1, efficient spectrum reuse will bring co-channel interference while improving spectrum efficiency. The quality of communication will be improved if co-channel interference can be suppressed effectively, especially the quality of service of the users on the cell edge. So the study of co-channel interference suppression is of great significance.Because of the lack of information of interference signals and the corresponding channels at the receiver, limited prior information must be fully tapped to solve the co-channel interference suppression problem. What is to be studied in this paper is to eliminate/suppress the interference signal from the received mixed signals without the information of interference signals and the channels.After analyzing the characteristics of interference signals in LTE system, the interference suppression method based on the blind source separation was studied in this paper for synchronous and asynchronous interference, respectively. The main work is listed as follows:1. According to whether the structure of interference signals and the useful signals is aligned in time domain, the interference is divided into synchronous and asynchronous interference, respectively. Then the received mixed-signal can be modeled as linear instantaneous mixed-signal model and linear convolution mixed-signal model.2. In the case of synchronous interference, effective blind source separation algorithm for complex mixed-signal separation was used to separate the received mixed-signal. Simulation results showed that useful signal could be separated from the mixed-signals and the interference signal from adjacent cells could be suppressed / eliminated effectively. The main features of this method is that only the statistical independence of the useful signal and interference signal is used, the structure properties of useful signal and interference signal and their corresponding channels are not needed.3. In the case of asynchronous interference, multi-channel blind deconvolution method for blind source separation was used to separate the useful signal and the co-channel interference. Then the useful modulation signal could be obtained. This method works well when the signal to noise ratio is not low, but the separation performance degrades when the SNR becomes low. Simulation results also shows that the separation method has significant performance advantages compared with the IRC method when the interference power is large.