基于微分动态规划的对地观测卫星条带姿态指令优化
Optimal Strip Attitude Command of Earth Observation Satellite using Differential Dynamic Programming
AI总结：
针对配备TDI相机的对地观测卫星条带成像最优扫描剖面问题，利用积分频率新增自由度，基于运动学推导角速度解析式，构建带约束最优控制问题并用DDP求解，可高效得到无收敛问题的近似最优解。
摘要：
本文针对配备时间延迟积分（TDI）相机的对地观测卫星（EOS），研究条带成像的最优扫描剖面问题。现代TDI相机可在成像过程中控制图像积分频率，为成像操作增加了一个额外自由度（DOF）。另一方面，现代敏捷型EOS能够对非平行地面目标成像，这类任务在运行期间需要较大的角速度与角加速度。我们利用这一自由度来最小化角速度等多种影响图像质量的因素。首先，我们基于运动学方程推导了角速度的解析表达式，随后利用这些表达式构建了带约束的最优控制问题（OCP），并采用微分动态规划（DDP）求解该问题。我们通过在多种实际场景中测试并与参考方法对比，验证了所提方法的有效性。仿真结果表明，该方法能够高效得到近似最优解，且不会出现不收敛问题。
英文摘要：
This paper addresses the optimal scan profile problem for strip imaging in an Earth observation satellite (EOS) equipped with a time-delay integration (TDI) camera. Modern TDI cameras can control image integration frequency during imaging operation, adding an additional degree of freedom (DOF) to the imaging operation. On the other hand, modern agile EOS is capable of imaging non-parallel ground targets, which require a substantial amount of angular velocity and angular acceleration during operation. We leverage this DOF to minimize various factors impacting image quality, such as angular velocity. Initially, we derive analytic expressions for angular velocity based on kinematic equations. These expressions are then used to formulate a constrained optimal control problem (OCP), which we solve using differential dynamic programming (DDP). We validate our approach through testing and comparison with reference methods across various practical scenarios. Simulation results demonstrate that our proposed method efficiently achieves near-optimal solutions without encountering non-convergence issues.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
