A Comprehensive Survey on Blind Source Separation for Wireless Adaptive Processing: Principles, Perspectives, Challenges and New Research Directions

With the rapid proliferation of wireless services, the frequency spectrum has become increasingly crowded, and the interferences and composite signals will be ubiquitous in the wireless receiver. For deeply dissecting and detecting the expected signals, the research community has to investigate the smart signal processing technology to resisting the influence of detrimental signals. For this purpose, blind source separation has been shown to be a promising method for achieving simultaneous spectrum utilization and wireless adaptive interference cancellation. The attractive features and appealing advantages of blind source separation make it an attractive theory for source extraction or recovery, which plays a crucial role in helping realize intelligent signal processing for wireless communication. It can recover the unobserved sources only from the wireless received mixed signals based on the features of the source signal exempted from channel estimation and synchronization manipulation. Wireless communication systems can benefit the high spectrum efficiency, strong anti-interference, and adaptive signal processing through the blind separation mechanism. So far, numerous researchers have made tremendous efforts to investigate this field for enhancing spectrum efficiency, anti-interference ability, and signal detection performance through employing the philosophy of blind separation. These meaningful and appealing research works motivate us to make a comprehensive survey with regard to this area. In this paper, the fundamental principle of the blind separation mechanism, involving independent component analysis, sparse component analysis, non-negative matrix factorization, and bounded component analysis will be reviewed briefly, and then, the critical technologies applied in various wireless communication systems will be overviewed, such as in direct-sequence code division multiplexing access, frequency hopping, orthogonal frequency-division multiplexing, multiple input multiple output, wireless sensor networks, cognitive radio networks, radio frequency identification devices, and communication security. In addition, the important research challenges and meaningful research directions pertaining to the area of blind separation applied in wireless communications systems are also discussed.