针对传统应急通信网规划方法难以适应复杂灾害环境的问题,提出一种基于改进神经网络的应急通信网规划方法.该方法融合多源异构数据,设计了图卷积长短期记忆(Graph Convolution Long Short-Term Memory,GC-LSTM)网络模型,实现对通信节点选址、链路规划、网络容量优化的智能化处理.在地震、洪水、台风等典型灾害场景下进实验验证,证实该方法在网络覆盖率、链路可靠性、容量匹配度等关键指标上均优于传统的图卷积神经网络(Graph Convolutional Network,GCN)和长短期记忆(Long Short-Term Memory,LSTM)网络模型,为提升应急通信网络的适应性和健壮性提供了新的技术途径.
Emergency Communication Network Planning Method Based on Improved Neural Network
Aiming at the problem that the traditional emergency communication network planning method is difficult to adapt to the complex disaster environment,an emergency communication network planning method based on improved neural network is proposed.This method integrates multi-source heterogeneous data,and designs a Graph Convolution Long Short-Term Memory(GC-LSTM)network model to realize intelligent processing of communication node location,link planning and network capacity optimization.Experiments in typical disaster scenarios such as earthquake,flood and typhoon show that this method is superior to the traditional Graph Convolutional Network(GCN)and Long Short-Term Memory(LSTM)network models in terms of network coverage,link reliability and capacity matching,which provides a new technical way to improve the adaptability and robustness of emergency communication networks.
emergency communication network planningGraph Convolution Long Short-Term Memory(GC-LSTM)networkdisaster emergency management
NETL
NSTL
万方数据
