Blind recognition of signals in LTE-U and Wi-Fi heterogeneous cognitive network

In order to reduce the pressure of licensed spectrum and improve the cellular network capacity, Long Term Evolution (LTE) in unlicensed spectrum (LTE-U) has attracted wide attention. However, there are many challenges to the heterogeneous network which is mainly formed by LTE-U and Wireless Fidelity (Wi-Fi) in the 5GHz unlicensed band. For improving the utilization of the unlicensed band, spectrum sharing of cognitive radio (CR) is an effective solution. For CR users (CRs) which have no information about primary users (PUs), the detected signal may be LTE or Wi-Fi signal, and how to accurately recognize both will be a challenge. In this paper, a blind recognition method of signals is proposed in LTE-U and Wi-Fi heterogeneous cognitive networks, and the signal model based on third Generation Partnership Project (3GPP) LTE and 802.11 Wi-Fi is considered. Without any prior information, the proposed method is based on the signal characteristic parameters which include the Fast Fourier Transformation (FFT) size, oversampling ratio and the number of effective subcarriers and these parameters are estimated based on autocorrelation and high order spectrum (HOS) properties of the signal. The blind recognition performance of the LTE and Wi-Fi signals is simulated versus different signal-to-noise ratio(SNR). Furthermore, simulation results show that the proposed method not only has better recognition performance, but also has lower complexity than the method based on eighth order statics and neural network classifier (EOS-NN).