Parameter Adaptive Sampling Inversion of Underwater Acoustic Go-back Channel Model Based on Bayes-MCMC
-
摘要: 高置信度水声双程信道仿真建模是水下目标回波仿真研究的基础内容, 对水下作战装备的研发具有重要作用。文中基于经典信道模型通过合理的假设建立了水声双程信道解析模型。在此基础上, 以Bayes-MCMC反演算法为核心, 分析水声双程信道参数反演问题特点, 设计了Metropolis-Hastings自适应单维度串行采样算法, 实现了基于回波信号的信道模型参数高效反演。经仿真及实测数据验证, 所设计的自适应采样反演方法其结果具有较好的一致性和收敛性, 在水下作战装备仿真测试方面具有良好的工程应用前景。
-
关键词:
- 水下作战装备 /
- 水声双程信道 /
- Bayes-MCMC /
- 回波信号反演
Abstract: High-confidence underwater acoustic go-back channel modeling is an essential part of the study of target echo simulation and plays an important role in the development of underwater operation equipment. Based on the classical channel model and reasonable assumptions, an analytical model of an underwater acoustic go-back channel is established. Using the Bayes-MCMC inversion algorithm as the core, the characteristics of the inversion problem of underwater acoustic channel parameters were analyzed, and the Metropolis-Hastings adaptive single-dimension serial sampling algorithm was designed to realize efficient channel model parameter inversion based on echo signals. The results of the simulation and measured data show that the proposed adaptive sampling inversion method has good consistency and convergence and has good engineering application prospects in underwater operation equipment simulation tests. -
图 1 水声双程信道反演算法基本流程
Figure 1. Basic process of underwater acoustic go-back chan- nel inversion algorithm
图 2 MCMC采样结果及其与真值分布对比
Figure 2. MCMC Sampling results and comparison with their true distributions
图 3 模型参数采样过程及反演结果与真值分布对比
Figure 3. Sampling process of model parameter and comparison of inversion results with true distributions
图 5 不同SNR信号反演结果与真值分布对比
Figure 5. Comparison of inversion results with true distributions under different SNR signals
7 1~8组信号反演结果与真值分布对比
7. Comparison of inversion results with true distributions at 1th to 8th groups signals
图 8 1~8组反演参数变化趋势
Figure 8. Change trend of inversion parameters at 1th to 8th groups signals
图 9 第5、9组双程信道模型参数反演过程结果对比
Figure 9. Comparison of parameter inversion results of go-back channel model at 5th & 9th groups signals
图 10 第5、9组信号反演结果与真值分布对比
Figure 10. Comparison of inversion results with true distributions at the 5th & 9th groups signals
表 1 不同SNR信号参数反演结果对比
Table 1. Comparison of parameter inversion results under different SNR signals
SNR/dB $ \overline {{\sigma ^2}} $ $D({\sigma ^2})$ $ \overline K $ $D(K)$ 6 0.595 0.003 1.025 0.01 1 0.487 9.47×10−4 0.852 0.008 SNR/dB $ \overline {{D_K}{L_{{Z_{\text{2}}}}}} $ $ D(\overline {{D_K}{L_{{Z_{\text{2}}}}}} ) $ $ \overline {{{\tilde D}_{{Z_{\text{2}}}}}} $ $ D({\tilde D_{{Z_{\text{2}}}}}) $ 6 0.216 0.009 1.73% 1.07×10−4 1 0.322 0.0002 8.05% 5.5×10−5 
下载: 导出CSV
表 2 1~8组双程信道模型参数反演结果
Table 2. Parameter inversion results of go-back channel model at 1th to 8th groups signals
样本组数 $ \overline {{\sigma ^2}} $ $ \overline K $ $ \overline {{D_K}{L_{{Z_{\text{2}}}}}} $ $ \overline {{{\tilde D}_{{Z_{\text{2}}}}}} $ 1 0.313 1.833 0.878 3.2% 2 0.303 1.830 1.055 4.1% 3 0.302 1.853 0.832 3.4% 4 0.298 1.863 0.813 3.3% 5 0.338 1.814 0.901 3.8% 6 0.295 1.856 1.260 4.3% 7 0.304 1.825 0.933 3.5% 8 0.314 1.829 0.850 3.4%
下载: 导出CSV
表 3 第5、9组信号参数反演结果对比
Table 3. Comparison of inversion results at 5th & 9th groups signals
样本组数 $ \overline {{\sigma ^2}} $ $D({\sigma ^2})$ $ \overline K $ $D(K)$ 5 0.353 6.72×10−4 1.816 0.001 9 0.287 2.15×10−4 1.854 0.0004 样本组数 $ \overline {{D_K}{L_{{Z_{\text{2}}}}}} $ $ D(\overline {{D_K}{L_{{Z_{\text{2}}}}}} ) $ $ \overline {{{\tilde D}_{{Z_{\text{2}}}}}} $ $ D({\tilde D_{{Z_{\text{2}}}}}) $ 5 0.901 0.117 3.83% 9.19×10−5 9 0.547 0.069 2.65% 7.67×10−5
下载: 导出CSV
-
[1] 徐瑜, 倪小清, 夏红梅, 等. 蛙人探测声呐发展现状及关键技术[J]. 舰船电子工程, 2017, 37(3): 1-3,126. doi: 10.3969/j.issn.1672-9730.2017.03.001Xu Yu, Ni Xiao-qing, Xia Hong-mei, et al. Current Status of Diver Detection Sonar and Key Technologies[J]. Ship Electronic Engineering, 2017, 37(3): 1-3,126. doi: 10.3969/j.issn.1672-9730.2017.03.001 [2] 邓新文, 叶浩亮. UUV声纳浅海声信道频率特性分析[J]. 鱼雷技术, 2013, 21(3): 184-188.Deng Xin-wen, Ye Hao-liang. Frequency Characteristics Analysis of Shallow Water Acoustic Channel for UUV Sonar[J]. Torpedo Technology, 2013, 21(3): 184-188. [3] 管日升, 范新刚, 杨文瑞. 基于主动时反技术的水下目标探测仿真[J]. 舰船科学技术, 2020, 42(17): 135-139. doi: 10.3404/j.issn.1672-7649.2020.09.026Guan Ri-sheng, Fan Xin-gang, Yang Wen-rui. Simulation of Underwater Target Detection Based on Active Time Reversal[J]. Ship Science and Technology, 2020, 42(17): 135-139. doi: 10.3404/j.issn.1672-7649.2020.09.026 [4] 李娜, 郝程鹏, 施博, 等. 水下修正空时自适应检测的性能分析[J]. 水下无人系统学报, 2018, 26(2): 133-139.Li Na, Hao Cheng-peng, Shi Bo, et al. Performance Analysis of Underwater Modified Space-Time Adaptive Detec- tion[J]. Journal of Unmanned Undersea Systems, 2018, 26(2): 133-139. [5] 王奇, 王英民, 诸国磊. 基于特征提取的匹配场处理[J]. 火力与指挥控制, 2014, 39(5): 97-100,106. doi: 10.3969/j.issn.1002-0640.2014.05.024Wang Qi, Wang Ying-min, Zhu Guo-lei. Matched Field Processing Based on Feature Extraction[J]. Fire Control & Command Control, 2014, 39(5): 97-100,106. doi: 10.3969/j.issn.1002-0640.2014.05.024 [6] 王海燕, 姜喆, 申晓红, 等. 两种水声信道模型中信噪比估计算法[J]. 西北工业大学学报, 2009, 27(3): 368-372. doi: 10.3969/j.issn.1000-2758.2009.03.016Wang Hai-yan, Jiang Zhe, Shen Xiao-hong, et al. Modifying SNR-Independent Velocity Estimation Method to Make it Suitable for SNR Estimation in Shallow Water Acoustic Communication[J]. Journal of Northwestern Polytechnical University, 2009, 27(3): 368-372. doi: 10.3969/j.issn.1000-2758.2009.03.016 [7] 董阳泽, 许肖梅, 刘平香, 等. 浅海声信道建模及其应用研究[J]. 系统仿真学报, 2010, 22(1): 47-50, 55. doi: 10.16182/j.cnki.joss.2010.01.032Dong Yang-ze, Xu Xiao-mei, Liu Ping-xiang, et al. Study on Modeling of Shallow Water Acoustic Channel and Its Applications[J]. Journal of System Simulation, 2010, 22(1): 47-50, 55. doi: 10.16182/j.cnki.joss.2010.01.032 [8] Radosevic A, Proakis JG, Stojanovic M. Statistical Characterization and Capacity of Shallow Water Acoustic Channels [C]//OCEANS 2009. US: IEEE, 2009. [9] Yang T C. Properties of Underwater Acoustic Communication Channels in Shallow Water[J]. The Journal of the Acoustical Society of America, 2012, 131(1): 231-236. [10] 许少华, 王皓, 王颖, 等. 一种改进的量子粒子群优化算法及其应用[J]. 计算机工程与应用, 2011, 47(20): 34-37. doi: 10.3778/j.issn.1002-8331.2011.20.010Xu Shao-hua, Wang Hao, Wang Ying, et al. Improved Quantum Particle Swarm Optimization Algorithm and Its Application[J]. Computer Engineering and Applications, 2011, 47(20): 34-37. doi: 10.3778/j.issn.1002-8331.2011.20.010 [11] Kim S M, Byun S H, Kim S G, et al. Experimental Analysis of Statistical Properties of Underwater Channel in a very Shallow Water Using Narrow and Broadband Signals[J]. OCEANS 2009. Biloxi, MS, United states, 2009 [12] Yardim C. Estimation of Radio Refractivity from Radar Clutter Using Bayesian Monte Carlo Analysis[J]. IEEE Transactions on Antennas And Propagation, 2006, 54(4): 1318-1327. doi: 10.1109/TAP.2006.872673 [13] Walsh B. Markov Chain Monte Carlo and Gibbs Sampling. Lexture Notes for EEB581[R]. [S.l.: s.n.], 2004. [14] 马天政, 吕昊, 张义民. 一种基于Bayes方法的随机模型修正方法[J]. 工程设计学报, 2016, 23(3): 206-211. doi: 10.3785/j.issn.1006-754X.2016.03.003Ma Tian-zheng, Lü Hao, Zhang Yi-min. Stochastic Model Updating Based on Bayesian Method[J]. Journal of Engineering Design, 2016, 23(3): 206-211. doi: 10.3785/j.issn.1006-754X.2016.03.003 [15] 高伟. 浅海混响统计反演海底声参数和界面统计参数研究[D]. 青岛: 中国海洋大学, 2008 [16] 盛峥, 黄思训, 曾国栋. 利用 Bayesian-MCMC 方法从雷达回波反演海洋波导[J]. 物理学报, 2009, 58(6): 4335-4341. doi: 10.3321/j.issn:1000-3290.2009.06.113Sheng Zheng, Huang Si-Xun, Zeng Guo-Dong. Ocean Duct Inversion from Radar Clutter Using Bayesian-Markov chain Monte Carlo Method[J]. Acta Physica Sinica, 2009, 58(6): 4335-4341. doi: 10.3321/j.issn:1000-3290.2009.06.113 [17] 王升超, 韩立国, 巩向博, 等. 基于局部采样MCMC方法的时移探地雷达反演[J]. 地球物理学报, 2022, 65(3): 1135-1143. doi: 10.6038/cjg2022P0201Wang Sheng-chao, Han Li-guo, Gong Xiang-bo, et al. Time-lapse Inversion of Ground Penetrating Radar (GPR) Based on Local Sampling MCMC Method[J]. Chinese Journal of Geophysics, 2022, 65(3): 1135-1143. doi: 10.6038/cjg2022P0201 -
点击查看大图
计量
- 文章访问数: 284
- HTML全文浏览量: 91
- PDF下载量: 36
- 被引次数: 0
