Abnormal transmission signal is the main reason for serious uplink transmission interference in ultra-dense networks, which leads to insufficient transmission interference suppression effect. Therefore, a local weighted regression based uplink transmission interference suppression for ultra-dense networks is proposed. Firstly, the robust local weighted regression is used to analyse the abnormal transmission signals of the uplink transmission link in the ultra-dense network, and to improved the suppression effect of the uplink transmission interference. Secondly, the adaptive time-frequency analysis is used to extract the characteristics of abnormal signals and determine the existence of uplink transmission interference. Finally, the residual neural network is used to identify the interference signal, and the interference reconstruction and suppression are combined to achieve the interference suppression of the uplink transmission in the ultra-dense network. Experimental results show that the reference signal received power gain, signal-to-noise ratio, and information average rate gain of the uplink station are optimised with the proposed interference suppression, and the maximum received power gain is up to 53.79 dB. Compared with the existing methods, the packet loss rate of the proposed method is significantly reduced, and the packet loss rate remains below 1%.
Sujuan Li (Thu,) studied this question.