The study investigates the trajectory estimation problem of a noncooperative gliding flight vehicle with complex and atypical maneuvers. An active switching multiple model (ASMM) method is proposed. This method employs a motion behavior model set (MBMS), a motion behavior recognition algorithm, and an active switching estimation and fusion algorithm. First, a recognizable MBMS, which can capture all the motion behaviors of a gliding flight vehicle, is established. Then, a motion behavior recognition algorithm based on recurrent neural networks (RNNs) is developed to obtain the current probability of each motion behavior. Then, an active switching estimation and fusion algorithm is proposed, in which the adopted models are actively chosen at each time instant according to a model selection strategy. Last, the proposed ASMM method is applied to a noncooperative gliding flight vehicle. The simulation results show that the proposed method has higher estimation precision and better dynamic performance.

该研究调查了具有非典型复杂动作的非合作滑行飞行器的轨迹估计问题。提出了一种主动切换多模型（ASMM）方法。该方法采用了运动行为模型集（MBMS），运动行为识别算法以及主动切换估计和融合算法。首先，建立了一个可识别的MBMS，它可以捕获滑行飞行器的所有运动行为。然后，开发了一种基于递归神经网络（RNN）的运动行为识别算法，以获得每种运动行为的当前概率。然后，提出了一种主动切换估计和融合算法，其中根据模型选择策略在每个时刻主动选择所采用的模型。持续，提出的ASMM方法被应用于非合作滑行飞行器。仿真结果表明，该方法具有较高的估计精度和较好的动态性能。