一种基于神经网络的700MHz干扰监控预警系统

A 700 MHz interference monitoring and warning system based on neural network

为降低700 MHz基站的广电信号干扰,提升网络性能,本文通过基于周期性长时段RB级干扰数据的干扰规避方案,将RB干扰划分为干扰区间,对干扰区间的总体干扰特征频度及相邻频段频度进行特征工程处理,将两种数据输入到训练后的神经网络模型,最终得到精准、可靠和稳定的干扰规避方案。通过复用干扰避让项目数据,实现了对某省700 MHz网络干扰情况的监控,可针对突发性干扰小区进行及时处理。

In order to reduce the radio and television signal interference of the 700 MHz base station and improve the network performance.This paper through the interference avoidance scheme based on periodic long period RB level interference data,divides RB interference into interference intervals,and then conducts feature engineering processing based on the overall interference characteristic frequency of the interference interval and the frequency of adjacent frequency bands,inputs two kinds of data into the trained neural network model,and fi nally obtains accurate,reliable and stable interference avoidance scheme.By reusing interference avoidance project data,monitoring the interference situation of the 700 MHz network in the province has been achieved,and sudden interference cells can be timely carried out.