| 引用本文： |
-
兰世玉,贾盈盈,丁良辉,等.一种面向无人集群抗干扰通信的链路预测与最优选择技术[J].电讯技术，2026，66（9）： - . [点击复制]
- LAN Shiyu,JIA Yingying,DING Lianghui,et al.Intelligent Prediction and Optimal Selection of Available Links for Unmanned Cluster Anti-jamming Communication[J].，2026，66（9）： - . [点击复制]
|
|
| 摘要: |
| 针对无人集群多信道通信环境中频谱感知效率低和信道切换频繁问题，提出了一种多信道干扰态势多步智能预测和最优链路多目标优化选择方法(Anti-jamming Link Prediction and Selection，AJ-LPS)。首先，面向多信道多时隙干扰功率预测难题，提出了融合长短期记忆(Long Short-term Memory，LSTM)网络长期记忆优势与KAN(Kolmogorov-Arnold Network)避免灾难性遗忘优势的LSTM-KAN预测网络，提升均方误差指标性能25.8%。然后，针对多时隙多信道链路选择难题，建立了最小化信道切换次数与最小化信道平均干扰功率的双目标优化模型，并提出分类排序遗传算法(Categorical Sorting Genetic Algorithm，CSGA)求解，获得比NSGA3(Non-dominated Sorting Genetic Algorithm III)等算法更优的求解性能。仿真结果表明，在多种扫频干扰与宽带干扰的叠加情况下，相对于单步干扰预测随机信道选择方法，提出的多步干扰预测双目标链路选择方法可减少信道切换次数26%，增加吞吐量8%，降低丢包率95%。 |
| 关键词: 无人集群 抗干扰通信 干扰预测 双目标优化 |
| DOI：10.20079/j.issn.1001-893x.250220002 |
|
| 基金项目:上海市重点实验室基金(STCSM 22DZ2229005) |
|
| Intelligent Prediction and Optimal Selection of Available Links for Unmanned Cluster Anti-jamming Communication |
| LAN Shiyu,JIA Yingying,DING Lianghui,YANG Feng,QIAN Liang |
| (School of Electronic Information and Electrical Engineering，Shanghai Jiao Tong University，Shanghai 200240，China) |
| Abstract: |
| For the problems of low spectrum sensing efficiency and frequent channel switching in unmanned clusters multi-channel communication environments，a multistep intelligent prediction of multi-channel interference situation and a multi-objective optimization selection method for optimal link is proposed.Firstly，for the difficult problem of multi-channel and multi-slot interference power prediction，the LSTM-KAN prediction network is proposed which fuses the long-term memory advantage of long short-term memory(LSTM) network and the advantage of Kolmogorov-Arnold Network(KAN) to avoid catastrophic forgetting，and improves the performance of mean square error by 25.8%.Then，for the multi-timeslot multi-channel link selection challenge，a dual-objective optimization model is established for minimizing the number of channel switches and minimizing the average channel interference power，and Categorical Sorting Genetic Algorithm(CSGA) is used to solve the problem，thus better performance than Non-dominated Sorting Genetic Algorithm Ⅲ(NSGA3) and other algorithms is obtained.Simulation results show that under the superposition of multiple swept interference and broadband interference，the proposed multistep interference prediction dual-target link selection method can reduce the number of channel switches by 26%，increase the throughput by 8%，and reduce the packet loss rate by 95%，compared with the single-step interference prediction random channel selection method. |
| Key words: unmanned cluster anti-jamming communication interference prediction dual-objective optimization |
《电讯技术》编辑部 