A Real-Time Detection Method and System for Antenna Cross Jumper Issue Based on Metric Learning

With the continuous expansion of network construction scale and the increasing number of mobile terminals, the issue of the cross jumper (CJ) in the base station antenna feeder system is becoming increasingly prominent. The cross jumper problem seriously affects network metrics and user experience. Traditional verification methods are time-consuming, labor-intensive, costly, and inefficient. To address this serious issue, we have designed an efficient and intelligent detection system and method. This method filters nearby users through dispatching Data, measures the signal strength characterization value on each antenna, and periodically collect and analyze those characterization value to determine whether there is cross jumper issue in the cell. We have constructed a Metric learning (ML) data analysis model based on Cluster Constraint - Metric Learning algorithm, to analyze feature data effectively. Through validation of operational network, our method can identify the cross jumper issue quickly and accurately, greatly improving customer satisfaction of telecommunications operators.