【摘要】 针对组合动力高超声速飞行器突防问题，提出了一种“双模态过载”的组合机动方式。基于经典微分对策理论解的符号函数的形式，结合组合动力高超声速飞行器的特点，提出了一种“小过载+大过载”的组合机动方式：机动方向由微分对策给出，机动幅值的大小则由模态给出。分析表明，“双模态”机动方式相比于“单模态”机动，在某些交战状态下能使脱靶量增大超过30%，可以有效改善突防效果，但脱靶量同时受到“大过载模态”持续时间和作用时机的影响；通过神经网络离线拟合脱靶量与“大过载模态”作用时机及持续时间的关系，能够根据交战态势信息在线生成大过载模态最优作用时机，取得了较好的拟合效果。结果表明，所提方法能够有效改善突防效果，且易于在线实现，有应用于工程实际的潜力。更多还原

【Abstract】 In order to solve the penetration problem of combined-powered hypersonic vehicles, a dualmodal overload maneuvering strategy is proposed. Based on the sign function form of the solution from classical differential game theory, and considering the characteristics of combined-powered hypersonic vehicles, a combined maneuver scheme using small and large G-leads is developed, in which the maneuvering direction is determined by the differential game solution, while the overload magnitude is governed by the selected maneuvering mode. Analysis indicates that, compared to the single-modal maneuver, the proposed dual-modal approach can increase the miss distance by more than 30% under certain engagement scenarios, thereby significantly improving penetration performance. However, the miss distance is also influenced by the duration and timing of the large overload mode. To address this, a neural network is employed to offline fit the relationship between the miss distance and the timing and duration of large overload mode, achieving accurate approximation. Based on engagement scenario information, the model can be adopted to generate the optimal timing of the large overload mode in real-time. Results demonstrate that the proposed method can effectively enhance penetration capability, and it is suitable for real-time implementation with potential for engineering applications.更多还原