基于时序知识图谱的军事通信场景认知与波形适变方法研究

Military communication scenario cognition and waveform adaptation based on temporal knowledge graph

摘要:

军事通信系统覆盖陆、海、空、天等作战全域,其面临的电磁环境、网络状态、业务需求等会随着作战行动的推进显著变化,因此如何认知瞬时的军事通信场景并给出自适应匹配的通信决策方案成为难题。针对场景驱动的自适应通信决策问题,研究了基于时序知识图谱的军事通信场景认知技术。首先,通过军事通信系统感知的多域信息构建军事通信场景本体架构,并从战场环境域、网络状态域以及用户行为域多维度对军事通信场景特征集进行了完整描述和特征表达。其次,对军事通信场景特征进行知识抽取、融合与表征,构建了面向军事通信场景的时序知识图谱。最后,在时序知识图谱的基础上嵌入知识表示学习模型,增强系统对未知通信场景的认知与推理预测能力,并对基于知识图谱的正交频分复用(OFDM)通信波形适变策略进行了功能验证。仿真实验表明,所提的军事通信场景认知与波形适变方法可以明显降低系统在城市、郊区、密林等不同场景下的误码率性能,有效提升了军事通信系统的自适应能力。

Abstract:

The military communication system covers land, sea, air and space combat do mains. The electromagnetic environment, network status, and communication requirements it faces undergo significant dynamic changes with the advancement of combat operations. Therefore, how to understand the instantaneous military communication scenarios and provide adaptive matching communication decision-making solutions has become a challenge. To address the scenario driven adaptive decision-making problems, this paper investigated the military communication scenario cognition technology based on temporal knowledge graph. Firstly, a military communication scenario ontology architecture was constructed based on the multi-domain information from the military communication system, and a comprehensive description and feature expression of the military communication scenario feature set were provided from multiple dimensions such as battlefield environment domain, network status domain, and user behavior domain. Secondly, knowledge extraction, fusion, and representation of military communication scenario feature were conducted to build a temporal knowledge graph for military communication scenarios. Finally, a knowledge representation learning model was embedded on the basis of the temporal knowledge graph to enhance the system’s ability to cognize and reason about unknown communication scenarios. The functionality of the orthogonal frequency division multiplexing (OFDM) communication waveform adaptation strategy based on the knowledge graph was verified. Simulation results show that the proposed military communication scenario cognition and waveform adaptation method can significantly reduce the bit error rate of the system in different scenarios such as cities, suburbs and dense forests, effectively improving adaptability of the system.