Abnormal signal detection in grid monitoring system based on overcomplete-basis independent component analysis

The higher capacity requirement for 5G mobile communication network poses significant technical challenge. Ultra-dense network (UDN) is believed to be one of the promising solutions. However, current wireless signal monitoring system is no longer capable of monitoring signals from massive base stations and detecting abnormal signals in UDN. In this paper, an innovative grid-monitoring strategy is presented to meet the need. In addition, a blind source separating method based on overcomplete-basis independent component analysis (OB-ICA) is verified as a preferable method to detect abnormal signals in the grid-monitoring system. Simulation on a MATLAB platform shows that OB-ICA method obtains 5% to 10% improvements in detection rate with respect to classic ICA method, which confirms that OB-ICA method is a competitive solution for detecting abnormal signals. Moreover, an optimized layout of monitoring stations corresponding to the location of base stations is proposed in this paper.