基于现实网络数据的通信感知一体化网络覆盖预测与优化
A Real-World Data-Based Network Coverage Prediction and Optimization for Integrated Sensing and Communications
李昕昊 1韩凯峰 2朱光旭1
作者信息
- 1. 香港中文大学(深圳),中国 深圳 518172;深圳市大数据研究院,中国 深圳 518172
- 2. 中国信息通信研究院,中国 北京 100083
- 折叠
摘要
传统基于人工经验、试错迭代的低效方法已无法适应未来多功能网络的参数优化.为了提升通信感知一体化(ISAC)网络的优化效率以及服务质量,亟需建立系统级的全局网络性能建模与优化新范式.为此,提出了一种面向6G的ISAC网络覆盖预测与优化框架.该框架利用离线的现实网络数据与多波束信道建模的方法,实现任意天线参数下的ISAC网络性能预测,并且利用零阶块坐标下降等数学优化工具求解黑盒优化问题,实现网络参数的精准寻优.基于所提框架,我们对现实世界中的空地协同ISAC网络进行了系统级性能仿真,并对低空用户进行感知性能优化,对地面用户进行通信性能优化.基于真实现网数据的实验结果表明,所提方法在性能上显著超越了传统的建模与优化方法.
Abstract
The conventional network optimization methods,based on manual experience and trial-and-error iterations are no longer suitable for parameter optimization of future multifunctional networks.To enhance the optimization efficiency and service quality of integrated sens-ing and communication(ISAC)networks,there is an urgent need to establish a new paradigm for system-level network performance model-ing and optimization.To this end,a 6G-oriented ISAC network coverage prediction and optimization framework is proposed.This framework utilizes offline real network data and multi-beam channel modeling methods to predict ISAC network performance under various parameters and employs advanced mathematical optimization tools such as zero-order block coordinate descent to solve black-box optimization prob-lems,achieving precise network parameter optimization.Based on the proposed framework,we conducted system-level performance simu-lations for real-world aerial-ground cooperative ISAC networks,optimizing sensing performance for low-altitude users and communication performance for ground users.Experimental results demonstrate that our method significantly outperforms traditional modeling and optimi-zation methods.
关键词
网络智能优化技术/通信感知一体化/本地化信道估计/覆盖预测/零阶优化/低空经济Key words
network optimization/integrated sensing and communication(ISAC)/localized statistical channel estimation/coverage prediction/zero-order optimization/low-altitude economy引用本文复制引用
基金项目
第八届中国科协青年人才托举工程项目(2022QNRC001)
广东省基础与应用基础研究重大项目(2023B0303000001)
国家自然科学基金面上项目(62371313)
广东省基础与应用基础研究基金面上项目(2022A1515010109)
深圳市龙岗区科技创新专项资金项目(LGKCSDPT2023002)
出版年
2024
NETL
NSTL
万方数据