基于机动动作元的敌机战术机动在线识别方法

敌机战术机动动作的在线识别可为空战态势感知、任务规划和辅助决策提供重要参考信息.针对雷达数据不能提供敌机姿态信息，导致战术机动动作难以直接描述的问题，提出了一种基于空战机动动作元的两级识别方法.首先分析了雷达传感器提供的特征信息；其次根据战术动作的航迹曲线，将空战战术动作分解为11种基本的机动动作元，降低了动作描述的复杂度；然后采用随机森林和支持向量机设计了两级识别方法，层级1利用航迹参数识别机动动作元，层级2根据动作元序列和运动参数识别战术动作.通过对仿真飞行数据进行验证，表明机动动作元能有效描述典型机动动作，提出的方法能够实现敌机战术机动的准确在线识别. 

Online Identification Method for Tactical Maneuver of Target Based on Air Combat Maneuver Element

The online identification for tactical maneuver of target can provide important information for air combat situation sensing, mission planning and aid decision making. Since the radar data cannot provide the posture information, it is difficult to describe the tactical maneuver of the enemy vehicle. In this paper, a two-stage recognition system for tactical maneuver of target based on air combat maneuver element (ACME) was proposed. First, the information from airborne radar was analyzed. Then, tactical maneuver was decomposed into eleven ACMEs according to the trajectory analyzing, thus simplifying the maneuver representation. And then, a recognition method utilizing random forest and support vector machine was designed, in the first stage, recognizing ACMEs with trajectory parameters, in the second stage, identifying tactical maneuver based on ACMEs sequence and motion parameters. Finally, the simulation results show that typical maneuver can be represented effectively by ACMEs and identified correctly by the recognition method.