自动化飞行训练评估中的战机机动动作识别

针对战机飞行员自动化飞行训练评估对于机动动作的在线识别需求，提出了一种改进的基于动态贝叶斯网络的机动动作识别方法。首先，分析了仪表、简单特技和复杂特技飞行科目的机动动作特征。然后，根据战机飞行过程中机动动作与特征参数的因果关系，建立了机动动作识别动态贝叶斯网络模型，克服了传统方法需要滚转角信息，在实际飞行训练中难以通过雷达探测实时获取的缺点。同时，通过设计模型在线调用机制，有效降低了计算复杂度。实验结果表明，所提方法对于战机机动动作识别率高、实时性好，能够满足在线应用需求。 

Recognition of fighter maneuver in automatic flight training evaluation

An improved online recognition method for fighter maneuver based on dynamic Bayesian network is proposed for automatic flight training evaluation. First, the maneuver characteristics of instrument, simple stunt and complex stunt flight are analyzed. Then, according to the causal relationship between maneuver and characteristic parameters during flight process of fighter, a dynamic Bayesian network model for maneuver recognition is established, which overcomes the shortcomings of traditional methods, such as the need for roll angle information which is difficultly obtained in real time through radar detection in actual flight training. At the same time, the computational complexity is reduced by designing the online invocation mechanism of the model. Experimental results show that this method has high fighter maneuver recognition rate and good real-time performance, and can meet the needs of online application.