Finding Probability of Submarine-Launched Acoustic Homing Torpedo Based on Gaussian Process Regression
-
摘要: 明确潜射声自导鱼雷的发现概率, 对相关战术制定具有显著作用。传统解析算法和统计算法无法平衡概率评估的快速性和精确性, 针对此问题, 文中提出了一种基于高斯过程回归的发现概率评估模型, 以及基于解析模型的训练数据集生成方法, 并在特定态势下开展了发现概率评估的数值仿真。结果显示, 文中所提方法具有很好的评估效果, 可为相关战场决策提供理论支撑。Abstract: The determination of the probability associated with submarine-launched acoustic homing torpedoes significantly influences tactical formulation. Conventional analytical and statistical algorithms encounter challenges in balancing the trade-off between the speed and precision of probability assessment. In response to this issue, this paper introduces a novel estimation model for assessing the probability of detection based on Gaussian process regression. Additionally, a methodology is proposed for generating training data using an analytical model. Subsequently, numerical simulations were conducted within a specific battlefield scenario. The outcomes illustrate the superior performance of the proposed method, offering valuable theoretical insights for informed decision-making in relevant battlefield contexts.
-
表 1 战场态势参数设置
Table 1. Parameters setting of battlefield situation
名称 参数 初始雷目距离/km 10∶5∶50 初始目标舷角/(°) 0∶10∶180 鱼雷速度/kn 50 声自导探测距离/m 1500 声自导探测扇面半角/(°) 50 鱼雷初始段航程/m 500 鱼雷回转半径/m 50 鱼雷初始发射提前角/(°) 20 目标速度/kn 20 目标航向/(°) 180 鱼雷速度误差标准差/kn 3 鱼雷航向误差标准差/(°) 1 探测目标速度误差标准差/kn 3 探测目标航向误差标准差/(°) 1 目标速度误差标准差/kn 3 目标航向误差标准差/(°) 1 目标初始方位探测误差标准差/(°) 1 目标初始距离探测误差标准差/m 1%初始距离 表 2 不同核函数均方根误差
Table 2. RMSE of different kernel function
高斯核函数 RMSE RQK 0.010 640 SEK 0.011 284 MK5/2 0.010 881 EK 0.012 799 表 3 需预测发现概率的新态势条件
Table 3. New situation for which the finding probability is needed to be predicted
编号 初始雷目距离/m 初始目标舷角/(°) 1 12 345 0~180(间隔1°) 2 34 567 0~180(间隔1°) 表 4 预测和验证相对误差范围及态势占比
Table 4. Relative error range of prediction and validation and percentage of battlefield situation
类别 相对误差范围 态势占比 预测 0%~1% 97.24% 0%~2% 100% 验证 0%~3% 79.83% 0%~5% 93.92% -
[1] 孟庆玉, 张静远, 宋保维. 鱼雷作战效能分析[M]. 北京: 国防工业出版社, 2003. [2] 杨绪升, 刘建兵, 周庆飞. 声自导鱼雷射击诸元及误差对其捕获概率的影响[J]. 指挥控制与仿真, 2009, 31(5): 93-97. doi: 10.3969/j.issn.1673-3819.2009.05.024Yang Xusheng, Liu Jianbing, Zhou Qingfei. Effect on acoustic homing torpedo capture probability of fire elements and errors[J]. Command Control & Simulation, 2009, 31(5): 93-97. doi: 10.3969/j.issn.1673-3819.2009.05.024 [3] 薄玉成. 武器系统设计理论[M]. 北京: 北京理工大学出版社, 2010. [4] 谢超, 周景军, 万亚民, 等. 基于脱靶量散布的超空泡鱼雷命中概率研究[J]. 水下无人系统学报, 2022, 30(2): 237-244,253. doi: 10.11993/j.issn.2096-3920.2022.02.015Xie Chao, Zhou Jingjun, Wan Yamin, et al. Research on hitting probability of supercavitating torpedo based on dispersion of miss distance[J]. Journal of Unmanned Undersea Systems, 2022, 30(2): 237-244,253. doi: 10.11993/j.issn.2096-3920.2022.02.015 [5] 谢超, 周景军, 万亚民, 等. 超空泡鱼雷命中概率解析方法[J]. 水下无人系统学报, 2022, 30(5): 656-664. doi: 10.11993/j.issn.2096-3920.202111002Xie Chao, Zhou Jingjun, Wan Yamin, et al. Analytical method for hitting probability of supercavitating torpedo[J]. Journal of Unmanned Undersea Systems, 2022, 30(5): 656-664. doi: 10.11993/j.issn.2096-3920.202111002 [6] 李勐, 代志恒. 直航鱼雷发现概率的解析计算方法[J]. 指挥控制与仿真, 2007, 39(4): 55-59.Li Meng, Dai Zhiheng. Analytic method for calculating probability of direct torpedo[J]. Commend Control & Simulation, 2007, 39(4): 55-59. [7] 武志东, 于雪泳, 许兆鹏. 潜射鱼雷命中概率的解析计算 通式及应用[J]. 水下无人系统学报, 2021, 29(2): 203-209.Wu Zhidong, Yu Xueyong, Xu Zhaopeng. Analytic formula and employment of the hitting probability for sub-launched torpedo[J]. Journal of Unmanned Undersea Systems, 2021, 29(2): 203-209. [8] 胡星志, 王旭, 江雄, 等. 基于高斯过程回归的高超声速飞行器不确定轨迹预测[J]. 空天技术, 2022(4): 49-61.Hu Xingzhi, Wang Xu, Jiang Xiong, et al. Uncertain trajectory prediction of hypersonic flight vehicles based on Gaussian process regression[J]. Aerospace Technology, 2022(4): 49-61. [9] Quinonero-Candela J, Rasmussen C E. A unifying view of sparse approximate Gaussian process regression[J]. The Journal of Machine Learning Research, 2005, 6: 1939-1959. [10] Schulz E, Speekenbrink M, Krause A. A tutorial on Gaussian process regression: Modelling, exploring, and exploiting functions[J]. Journal of Mathematical Psychology, 2018, 85: 1-16. doi: 10.1016/j.jmp.2018.03.001