位置:成果数据库 > 期刊 > 期刊详情页
基于Boosting学习算法的雷达弹道识别
  • 期刊名称:弹箭与制导学报
  • 时间:0
  • 页码:193-196
  • 分类:TJ012[兵器科学与技术—兵器发射理论与技术]
  • 作者机构:[1]解放军炮兵学院,合肥230031, [2]中国科学院自动化研究所,北京100190
  • 相关基金:基金项目:国家自然科学基金(60835002.60975040)资助
  • 相关项目:基于损失函数的统计机器学习算法及其应用研究
中文摘要:

弹道外推技术在炮位雷达的侦察和校射中起关键作用,弹道外推的精度直接决定着炮位侦察校射雷达的性能。在文献[1]中,作者提出了将弹道外推分为弹道识别和特定弹道外推两个阶段,并用支持向量机方法对弹道识别进行了系统研究。文中引进Boosting学习算法进行弹道识别。仿真结果表明,基于决策树的Boosting学习算法是一种有效的弹道识别方法,并且识别精度高于基于核技巧的支持向量机方法。

英文摘要:

Trajectory prediction plays a crucial role in reconnaissance and adjustment of radar, and the performance of radar for re- connaissance and adjustment is directly determined by its accuracy. In our paper [1], it was proposed that the stage of trajectory prediction can be divided into the recognition phase and prediction phase of specific trajectories, and the application of SVM in trajectory recognition was systematically investigated. In this paper, the Boosting classification technique was introduced to recognize the trajectories. Several experiments indicate that the efficient decision-tree-based Boosting algorithms reach higher precision than kernel-based SVM.

同期刊论文项目
期刊论文 30 会议论文 56
同项目期刊论文