-
摘要: 在低軌衛星網絡中，衛星運行速度快、運行周期較短，星間鏈路動態變化。為了及時感知星間鏈路狀態并選擇正確的路由，提出一種基于樹突神經網絡的低軌衛星智能感知路由算法，通過衛星之間的可視性約束分析星間建鏈情況，實現星間鏈路態勢感知；通過實時構造訓練集，利用樹突神經網絡自動調整全局衛星網絡鏈路的權值，進而優化傳統迪杰斯特拉(Dijkstra)算法，實現星間鏈路質量感知，給出智能路由決策；通過周期性監測衛星網絡拓撲，實時修正初始路由路徑。仿真結果表明，基于樹突神經網絡的路由算法復雜度低，路徑時延、時延抖動及丟包率均低于傳統啟發式路由算法和Dijkstra路由算法。Abstract: In a low-Earth orbit (LEO) satellite network, the satellite operation speed is high, the operation cycle is short, and intersatellite links change dynamically. To sense the intersatellite link state in time and select the correct route for an intelligent routing decision, a dendritic network-based intelligent-aware routing algorithm for LEO satellites is proposed in this paper. This algorithm divides the intersatellite link routing of an LEO satellite network into situation-aware, quality-aware, and routing-decision stages and establishes a routing policy framework with real-time correction capability from the source node to the destination. This approach overcomes the problems of the limited selection of routing paths from fixed labels of existing deep learning-based routing algorithms and the long convergence time of reinforcement learning-based routing algorithms.In the intersatellite link situational awareness stage, the intersatellite visibility of the entire LEO satellite network is periodically obtained by analyzing the constraint conditions of the intersatellite link establishment. In the intersatellite link quality perception stage, the final output of the probabilistic forwarding matrix based on the ant colony algorithm is used as the label of the training set, and the corresponding intersatellite link quality is evaluated using the probability value of the current node by selecting the next hop node. By changing the weight coefficients in the path cost function under different load states, more effective training set label data can be collected, which can be consequently used to improve the performance of the trained dendritic network. Moreover, the training set can be optimized in real-time through semi-supervised learning. The trained dendritic network is used to analyze and process the link state parameters, perceive the comprehensive service quality of the link, and output the evaluation value matrix of the next hop routing. It is also used to automatically adjust the weight of the global satellite network link. Meanwhile, the traditional Dijkstra algorithm is optimized to realize the quality perception of the intersatellite link. In the routing decision stage, the reciprocal of the evaluation value matrix is used as the adjacency matrix to pass the shortest-path algorithm. Then, the initial routing path between the source and destination nodes is obtained. Finally, the initial path is corrected via periodic monitoring to cope with the failure of the satellite node. The simulation results show that the routing algorithm based on the dendritic network has low computational complexity and fast convergence. The algorithm can determine the status of the intersatellite link establishment in time, assess the quality of the intersatellite link in real-time, and automatically avoid congested satellite nodes. Accordingly, its end-to-end path delay, delay jitter, and packet loss rate are lower than those of the traditional heuristic routing algorithm and Dijkstra routing algorithm.
-
圖 4 某時刻衛星相對位置示意圖
Figure 4. Image of the relative position of the satellite at a certain moment
表 1 Walker參數設置
Table 1. Parameter setting of the Walker constellation
Type Number of sats
per planeNumber of planes Interplane spacing RAAN
spreadDelta 8 8 1 360° 
下載: 導出CSV
表 2 鏈路參數設置
Table 2. Parameter setting of the link
${\rm{Dela}}{{\rm{y}}_{{\rm{max}}} }$/ ms ${\rm{Los}}{{\rm{s}}_{{\rm{max}}} }$ ${\rm{Jitte}}{{\rm{r}}_{{\rm{max}}} }$ 100 0.0005 ${\text{3} }{\text{.6} } \times {\text{1} }{ {\text{0} }^{ {{ - 4} } } }$
下載: 導出CSV
www.77susu.com<span id="fpn9h"><noframes id="fpn9h"><span id="fpn9h"></span> <span id="fpn9h"><noframes id="fpn9h"> <th id="fpn9h"></th> <strike id="fpn9h"><noframes id="fpn9h"><strike id="fpn9h"></strike> <th id="fpn9h"><noframes id="fpn9h"> <span id="fpn9h"><video id="fpn9h"></video></span> <ruby id="fpn9h"></ruby> <strike id="fpn9h"><noframes id="fpn9h"><span id="fpn9h"></span> -
參考文獻
[1] Peng J, Sun M Y, Teng X Q. 6G vision and application scenario outlook. China Ind Inf Technol, 2020(9): 18 doi: 10.3969/j.issn.1674-9138.2020.09.003彭健, 孫美玉, 滕學強. 6G愿景及應用場景展望. 中國工業和信息化, 2020(9):18 doi: 10.3969/j.issn.1674-9138.2020.09.003 [2] Wu S G, Wang H Y, Wang Y, et al. Research and analysis of low-orbit satellite network routing technology. Satellites Networks, 2021(9): 66吳署光, 王宏艷, 王宇, 等. 低軌衛星網絡路由技術研究分析. 衛星與網絡, 2021(9):66 [3] Shao F W. Research on the Link Route Technology of the Inter-satellite Link Network of the Satellite Navigation System [Dissertation]. Beijing: University of Chinese Academy of Sciences, 2017邵豐偉. 衛星導航系統星間鏈路網絡建鏈路由技術研究[學位論文]. 北京: 中國科學院大學, 2017 [4] Huang J H. Research on Optimal Design Technology of Inter-Satellite Link Networks in Global Navigation Satellite Systems [Dissertation]. Changsha: National University of Defense Technology, 2018黃今輝. 衛星導航系統星間鏈路網絡優化設計技術研究[學位論文]. 長沙: 國防科技大學, 2018 [5] Hao L C, Ren P Y, Du Q H. Satellite QoS routing algorithm based on energy aware and load balancing // International Conference on Wireless Communications and Signal Processing. Nanjing, 2020: 685 [6] Gao H, Wang L, Huang W D, et al. Routing optimization method for fast return of data on overseas satellites in Beidou Global Navigation Satellite System. Chin Space Sci Technol, 2018, 38(2): 9 doi: 10.16708/j.cnki.1000-758x.2018.0002高賀, 王玲, 黃文德, 等. 北斗全球衛星導航系統境外星數據快速回傳的路由優化方法. 中國空間科學技術, 2018, 38(2):9 doi: 10.16708/j.cnki.1000-758x.2018.0002 [7] Liao G Y. Research on Routing Algorithm of Low-orbit Satellite Network Based on Inter-Satellite Link [Dissertation]. Chongqing: Chongqing University of Posts and Telecommunications, 2020廖光燕. 基于星間鏈路的低軌衛星網絡路由算法研究[學位論文]. 重慶: 重慶郵電大學, 2020 [8] Li X T, Zhang Y S. Artificial intelligence routing method suitable for SDN network suitable for LEO satellites. Electron Meas Technol, 2020, 43(22): 109 doi: 10.19651/j.cnki.emt.2005049李新桐, 張亞生. 一種適用于低軌衛星的SDN網絡人工智能路由方法. 電子測量技術, 2020, 43(22):109 doi: 10.19651/j.cnki.emt.2005049 [9] Liu H Y, Sun F C, Li H B, et al. A satellite network QoS routing scheme based on machine learning. J Central South Univ Sci Technol, 2013, 44 (Suppl 2): 263劉賀語, 孫富春, 李洪波, 等. 一種基于機器學習的衛星網絡QoS路由機制. 中南大學學報(自然科學版), 2013, 44 (增刊2): 263 [10] Wang X T, Dai Z Q, Xu Z. LEO satellite network routing algorithm based on reinforcement learning // 2021 IEEE 4th International Conference on Electronics Technology. Chengdu, 2021: 1105 [11] Tu Z, Zhou H C, Li K, et al. A routing optimization method for software-defined SGIN based on deep reinforcement learning // 2019 IEEE Globecom Workshops. Waikoloa, 2019: 1 [12] Yang Y, Lv G H, Zhao H, et al. Survey on deep learning applicatons in software defined networking research. J Softw, 2020, 31(7): 2184 doi: 10.13328/j.cnki.jos.006039楊洋, 呂光宏, 趙會, 等. 深度學習在軟件定義網絡研究中的應用綜述. 軟件學報, 2020, 31(7):2184 doi: 10.13328/j.cnki.jos.006039 [13] Gao H S, Tang X, Cao W B. Loop-free routing algorithm based on deep neural network. Radio Eng, 2022, 52(1): 101 doi: 10.3969/j.issn.1003-3106.2022.01.015高會生, 唐驍, 曹旺斌. 基于深度神經網絡的無環路由算法. 無線電工程, 2022, 52(1):101 doi: 10.3969/j.issn.1003-3106.2022.01.015 [14] Wu Y S. Deep reinforcement learning routing for dynamic topology networks. Telecommun Eng, 2021, 61(6): 659伍元勝. 面向動態拓撲網絡的深度強化學習路由技術. 電訊技術, 2021, 61(6):659 [15] Zhang L, Yan F, Zhang Y Y, et al. A routing algorithm based on link state information for LEO satellite networks // 2020 IEEE Globecom Workshops. Taipei, 2020: 1 [16] Liu X, Xie J S, Chen S W. Link quality aware routing scheme in LEO satellite network. Astronaut Syst Eng Technol, 2020, 4(2): 33劉洵, 謝金森, 陳雙武. 鏈路狀態感知的低軌衛星網絡路由機制. 宇航總體技術, 2020, 4(2):33 [17] Chen Y, Zhao L F, Liu H J, et al. Analysis of configuration and maintenance strategy of LEO walker constellation. J Astronaut, 2019, 40(11): 1296 doi: 10.3873/j.issn.1000-1328.2019.11.005陳雨, 趙靈峰, 劉會杰, 等. 低軌Walker星座構型演化及維持策略分析. 宇航學報, 2019, 40(11):1296 doi: 10.3873/j.issn.1000-1328.2019.11.005 [18] Wang W D, Wang C, Wang H W, et al. Dynamic cache allocation routing strategy of Internet of Things satellite node based on traffic prediction. J Commun, 2020, 41(2): 25 doi: 10.11959/j.issn.1000-436x.2020038王衛東, 王程, 王慧文, 等. 基于流量預測的物聯網衛星節點動態緩存分配路由策略. 通信學報, 2020, 41(2):25 doi: 10.11959/j.issn.1000-436x.2020038 [19] Zhang L, Yan F, Zhang Y Y, et al. LEO satellite routing algorithm based on inter-satellite link state information. Aerosp Shanghai Chin Engl, 2021, 38(4): 92張路, 燕鋒, 章躍躍, 等. 基于星間鏈路狀態的低軌衛星網絡路由算法. 上海航天(中英文), 2021, 38(4):92 [20] Li W, Sun L, Wang J Q, et al. Key technologies to enable 5G and TSN coordination for industrial automation. Chin J Eng, 2022, 44(6): 1044李衛, 孫雷, 王健全, 等. 面向工業自動化的5G與TSN協同關鍵技術. 工程科學學報, 2022, 44(6):1044 [21] Luo Z Y. Research of Routing Technology Based on Machine Learning in Integrated Satellite-Terrestrial Information Network [Dissertation]. Chengdu: University of Electronic Science and Technology of China, 2021羅澤耀. 天地一體化網絡下基于機器學習的路由技術研究[學位論文]. 成都: 電子科技大學, 2021 [22] Yin F J, Chu Q S. Research on QoS routing based on improved ant colony algorithm. J Liaoning Univ Nat Sci Ed, 2020, 47(4): 312 doi: 10.16197/j.cnki.lnunse.2020.04.005尹鳳杰, 褚群森. 基于改進蟻群算法的QoS路由研究. 遼寧大學學報(自然科學版), 2020, 47(4):312 doi: 10.16197/j.cnki.lnunse.2020.04.005 [23] Zhao J R, Liu J, Zhang R, et al. Optimization of QoS routing on LEO satellite network based on ant colony algorithm. Radio Commun Technol, 2021, 47(5): 590 doi: 10.3969/j.issn.1003-3114.2021.05.011趙晶蕊, 劉江, 張然, 等. 基于蟻群算法的LEO衛星網絡QoS優化路由. 無線電通信技術, 2021, 47(5):590 doi: 10.3969/j.issn.1003-3114.2021.05.011 [24] Liu G, Wang J. Dendrite net: A white-box module for classification, regression, and system identification. IEEE Trans Cybern,https://doi.org/10.1109/TCYB.2021.3124328 [25] Liu G. It may be time to improve the neuron of artificial neural network [J/OL]. TechRxiv preprint (2021-7-21) [2021-11-08].https://doi.org/10.36227/techrxiv.12477266 [26] Liu G, Wang J. A relation spectrum inheriting Taylor series: Muscle synergy and coupling for hand. Front Inform Technol Electron Eng, 2022, 23(1): 145 doi: 10.1631/FITEE.2000578 [27] Valadarsky A, Schapira M, Shahaf D, et al. Learning to route // Proceedings of the 16th ACM Workshop on Hot Topics in Networks. Hangzhou, 2017: 185 [28] Ma Z G, Xu X H, Liu X E. Three analytical frameworks of causal inference and their applications. Chin J Eng, 2022, 44(7): 1231馬忠貴, 徐曉晗, 劉雪兒. 因果推斷三種分析框架及其應用綜述. 工程科學學報, 2022, 44(7):1231 -
點擊查看大圖
圖(11) / 表(2)
計量
- 文章訪問數: 815
- HTML全文瀏覽量: 333
- PDF下載量: 120
- 被引次數: 0
