Study Of Blind Source Separation In Anti-interference Communication Systems

Currently, the spread spectrum communication technology and massive antenna systems have been widely used in wireless communication systems. There are increasing different kinds of complex interferences existing in wireless environment which make the performance of communication system worse. In order to improve the performance of spread spectrum communication systems and massive antenna systems, techniques of anti-interference must be applied. At present, in spread spectrum communication systems, most of the anti-interference methods are based on time-domain filtering and frequency-domain filtering. The anti-interference methods mainly based on LS and MMSE channel estimation algorithms, and ZF and MMSE decoding algorithms have been used to suppressing pilot contamination in massive antenna systems.Anti-interference technique based on blind source separation only uses received signal to separate source signal and complete the anti-interference. Now the widely used anti-interference method needs to know a priori information channel, when the channel information is unknown, the performance decreases quickly. Therefore, the study of blind source separation becomes very important in anti-interference communication systems.This paper proposes two new spread spectrum communication anti-interference algorithms based on blind source separation, as well as a new massive MIMO pilot contamination suppression algorithm based on blind source separation. In spread spectrum communications, the first one is anti-interference algorithm based on blind source separation in single channel. This algorithm regards the interference signal as a source, to separate the interference signal and spread spectrum signals by using independent component analysis which assisted by chip sequence. The proposed method is based on the whole statistical properties of the received signal. The second method is anti-interference algorithm based on space-time blind source separation in multi-channel. This algorithm uses the periodicity of the chip sequence to build the model of space-time blind source separation, makes the mixed matrix includes information of chip sequence and direction of arrival. Through the space-time blind source separation structure, the underdetermined model is reconstructed as an over-determined model. Then the independent component analysis algorithm which is assisted by chip sequence is used to completing interference suppression.In massive MIMO, this paper proposes a channel estimation algorithm and decoding algorithm based on blind source separation. Without the pilot training sequence, this algorithm uses the independent component analysis method which is assisted by the principal component analysis method to separate the received signal completely which overcomes the uncertainty of the separation sequence of independent component analysis method, and finally obtains user's channel estimation and data information.Theoretical analysis and simulation show that, in view of the spread spectrum communication systems, the single channel of anti-interference based on blind source separation algorithm is better than the existed time-domain filtering anti-interference algorithm based on bilateral LMS, frequency-domain adaptive filtering anti-interference algorithm. In multiple antenna spread spectrum system, the proposed anti-interference algorithm based on space-time blind source separation has better anti-interference performance than the traditional spatial-filtering anti-interference, space-time-filtering anti-interference method. For massive MIMO systems, the proposed decoding algorithm based on blind source separation has better performance than the widely used channel information unknown ZF decoding algorithm and MMSE decoding algorithm.