|
范学满(1989-),男,山东青岛人,博士,实习研究员,研究方向为机器学习、智能决策。 |
|
张 会(1971-),女,博士,副教授。 |
Copy editor: 许韦韦
收稿日期: 2018-09-25
修回日期: 2018-10-29
网络出版日期: 2022-05-16
Research on the Interference Effect Prediction of Anti-torpedo Decoys Based on AdaBoost
Received date: 2018-09-25
Revised date: 2018-10-29
Online published: 2022-05-16
鱼雷来袭时,潜艇通常通过发射诱饵和规避机动进行防御。根据本艇、诱饵和鱼雷的相对态势,实时、准确地预判诱饵的干扰效果即鱼雷能否发现本艇,对本艇进一步防御决策具有重要意义。目前,基于经验的预测无法保证准确率的要求,基于在线仿真的预测无法保证实时性要求。对此,采用机器学习将该问题转化为典型的二分类问题,以本艇、诱饵和鱼雷的相对态势作为分类特征,通过离线仿真生成训练数据集,以错误率降低剪枝决策树(Reduced Error Pruning Tree, REPTree)作为基分类器,构建了基于自适应增强(Adaptive Boosting,AdaBoost)的诱饵干扰效果预测模型。实验结果表明,模型具有良好的鲁棒性和准确性。
范学满 , 张会 . 基于AdaBoost的潜射防鱼雷诱饵干扰效果预测研究[J]. 指挥控制与仿真, 2019 , 41(3) : 52 -56 . DOI: 10.3969/j.issn.1673-3819.2019.03.011
Submarines usually launch baits and evade maneuver to intercept a torpedo. It is of great significance to predict the interference effect of decoys real time and accurately for the further defense decision-making of the submarines. At present, prediction methods based on experience cannot guarantee the accuracy, at the same time, prediction methods based on online simulation cannot meet the real-time requirements. In this regard, the machine learning method is used to transform the problem into binary classification problem. The relative situation of the submarine, acoustic decoy and torpedo is used as the classification features, and the REPTree (Reduced Error Pruning Tree) is used as basic classification algorithm, to construct a decoy’s interference effect prediction model based on AdaBoost (Adaptive Boosting). The experimental results show that the established AdaBoost prediction model has good robustness and accuracy.
Key words: operation assistant decision; decision tree; ensemble learning; AdaBoost
| [1] |
李斌, 王顺杰. 潜艇应用自航式声诱饵防御声自导鱼雷仿真研究[J]. 指挥控制与仿真, 2014, 36(3): 98-103.
|
| [2] |
王德志, 张孝顺, 刘前进, 等. 基于集成学习的孤岛微电网源--核协同频率控制[J]. 电力系统自动化, 2018, 42(10): 46-52.
|
| [3] |
|
| [4] |
陈法法, 杨晶晶, 肖文荣, 等. Adaboost-SVM集成模型的滚动轴承早期故障诊断[J]. 机械科学与技术, 2018, 37(2): 237-243.
|
| [5] |
贾科, 宣振文, 林瑶琦, 等. 基于Adaboost算法的并网光伏发电系统的孤岛检测法[J]. 电工技术学报, 2018, 33(5): 1106-1113.
|
| [6] |
杨叶梅, 陈新. 利用惯性传感器和AdaBoost算法的步态识别方法[J/OL]. 计算机应用研究, 2019, 36(5). [2018-03-09]. http://www.arocmag.com/article/02-2019-05-059.html
|
| [7] |
|
| [8] |
周志华. 机器学习[M]. 北京: 清华大学出版社, 2016: 66-69.
|
| [9] |
|
| [10] |
袁梅宇. 数据挖掘与机器学习WEKA应用技术与实践(第2版)[M]. 北京: 清华大学出版社, 2016: 1-4.
|
| [11] |
|
| [12] |
丁文哲, 李新洪, 杨虹. 基于AdaBoost的填充式防护结构超高速撞击损伤研究[[J/OL]]. 北京航空航天大学学报. http://doi.org/10.13700/j.bh.1001-5965.2018.0216
|
/
| 〈 |
|
〉 |