基于AI的天馈自优化实现研究

Research on Implementation of Self Optimization of Antenna Feed with AI

研究了下倾角动态迭代寻优算法、天馈最佳方向角定位算法和图像识别精灵算法。倾角寻优算法通过电子下倾方案迭代,遍历所有倾角组合,找出小区弱覆盖栅格占比、质差栅格占比与导频污染栅格占比综合得分最小的天馈倾角优化方案,与高斯分布模型算法得出的天馈最佳方向角,形成天馈调整脚本;进而驱动图像识别精灵机器人,实施网管远程天馈调整动作。构建天馈闭环管控系统,实施天馈自动化闭环评估优化。经验证,原有越区覆盖、质量得到有效改善。本方法快速、有效、便捷,适用于匹配5G权值快速优化的4G/5G网络覆盖质量联合优化。

It studies the dynamic iterative optimization algorithm of down tilt angle,the location algorithm of antenna and feed optimal direction angle and the image recognition algorithm. Through the iteration of the electronic down tilt scheme,the down tilt optimization algorithm traverses all the down tilt combinations and finds the antenna feed down tilt optimization scheme with the minimum comprehensive score of the cell weak coverage grid proportion,the poor quality grid proportion and the pilot pollution grid proportion,and then forms the antenna adjustment script with the best direction angle of the antenna obtained by the Gaussian distribution model algorithm,further more,combined with the image recognition wizard robot,it implements the remote antenna feed adjustment of network management. And the closed-loop evaluation are conducted to implement the closed-loop evaluation optimization of antenna feeder automaticly. It is verified that the coverage and quality of the original cross-district are improved effectively. This method is fast,effective and convenient,which is suitable for the joint optimization of 4 G/5 G network coverage quality.