Stalled redirection control of relaxed static stability aircraft

放宽静稳定度飞行器过失速重定向控制

School of Astronautics, Northwestern Polytechnical University, Xi'an 710072, China

Received: 21 May 2025

Abstract

Traditional aircraft face limitations in off-boresight maneuverability, making it challenging to effectively meet rear hemisphere target tracking requirements. This paper proposes a proximal policy optimization (PPO)-based intelligent control method to achieve aerodynamic-controlled post-stall maneuver redirection for relaxed static stability aircraft. First, the nonlinear aerodynamic characteristics of the aircraft at extremely high angles of attack are analyzed using the Jorgensen engineering method, revealing the impact of static instability on autorotation capability and segmenting the redirection process. Second, a Markov decision process (MDP) model is established to address nonlinear attitude control under ultra-high angles of attack, incorporating a potential energy-based reward function to guide the intelligent agent in learning optimal rudder deflection strategies. Finally, redirection simulations under low Mach number conditions demonstrate that aerodynamic control surfaces alone can enable rapid aircraft redirection while maintaining attitude stability during velocity zero-crossing phases. Monte Carlo experiments further validate the robustness of this control method under random initial conditions.

摘要

传统飞行器受离轴机动能力限制, 难以有效满足后半球目标跟踪需求。提出利用近端策略优化(PPO)智能控制方法实现放宽静稳定度飞行器在气动控制下的过失速机动重定向。基于Jorgensen工程方法分析了飞行器在超大攻角下的非线性气动特性, 揭示了静不稳定度对自翻转能力的影响, 并对重定向过程进行了分段划分。针对超大攻角非线性姿态控制问题进行了马尔可夫决策过程建模, 结合势能奖励函数设计引导智能体有效学习最优舵偏控制策略。在低马赫数条件下进行了重定向仿真实验, 结果表明仅依靠空气舵即可控制飞行器完成快速重定向, 并在速度过零段维持飞行器姿态稳定, 蒙特卡洛实验进一步验证了该控制方法在随机初始条件下的鲁棒性。