变体辅助的固定翼飞行器仿生栖落机动的控制方法研究

The bionic perching maneuver of fixed-wing aircraft is a new type of landing method without the need of runway. In the process of perching maneuver, fixed-wing aircraft emulate large or medium-sized birds, increase the angle of attack to the extent of deep stall, decelerate rapidly in the air, and finally land precisely at the prescribed location; and the aircraft dynamics is highly nonlinear, uncertain and under-actuated. The existing methods of perching maneuver control have narrow regions of attraction and cannot guarantee successful perching maneuver in the presence of wind disturbances. In this research project, considering the longitudinal motion of perching maneuvers, the morphing mechanism is utilized to enhance the controllability of aircraft during deep stall, and a new perching control design method, which guarantees large region of attraction and high disturbance rejection ability, is proposed. The aerodynamic model of perching maneuvers is constructed based on both flight experiment and aerodynamic theory. The dynamics of the morphing aircraft is modeled by using multi-body dynamics method. The control approaches with non-matching disturbance rejection ability for perching maneuvers aided by morphing are studied. A kind of trajectory and control integrated design strategy based on sums-of-square (SOS) method is proposed to enlarge the region of attraction. The concept of adaptive reference trajectory is proposed for the case of large disturbance, and the integrated design method for adaptive trajectory and tracking control is studied.

仿生栖落机动是固定翼飞行器的一种无需跑道的新颖降落方式。在仿生栖落机动过程中，固定翼飞行器模仿大中型鸟类，拉大迎角进入过失速状态，在空中快速减速，最终实现精准的降落；飞行器动力学具有高度非线性、不确定性、欠驱动与多体动力学特性。目前文献中的栖落机动控制方法吸引域小，且无法在风扰情况下实现成功的栖落机动。本项目针对栖落机动的纵向运动，采用变体部件辅助的方法改善过失速状态中飞行器的俯仰操纵性，在此基础上提出了一种具有大吸引域与高抗扰能力的新颖栖落机动控制设计方法。采用飞行实验与气动理论结合的方法建立栖落机动气动模型，采用多体动力学方法建立考虑风扰的变体飞行器动力学模型。研究具有抑制非匹配干扰能力的变体辅助栖落机动控制方法；以此为基础，基于平方和方法研究具有较大吸引域的轨迹与控制一体化设计策略；针对大扰动的情况，引入自适应参考轨迹的概念，研究自适应轨迹与跟踪控制律的一体化设计方法。

仿生栖落机动是固定翼飞行器的一种无需跑道的新颖降落方式。项目针对变体辅助的固定翼飞行器，设计了具有较大吸引域与高抗扰能力的纵向栖落机动控制律；提出了一套轨迹优化与跟踪控制的综合设计方法。对几类典型变体无人机的栖落机动纵向运动进行了气动建模与动力学建模。考虑外部风扰动，研究基于模型预测控制的栖落机动跟踪控制问题。针对变体无人机的栖落机动，研究多项式模糊控制器设计方法，并探讨了不同变体方案对栖落机动性能的影响。为了进一步提高扰动抑制能力，研究了在风扰下无人机的栖落机动的强化学习控制策略设计。为了扩大初始容许误差范围，研究了栖落机动的扩大吸引域的轨迹跟踪控制方法。基于迭代学习模型预测控制策略，研究了无人机栖落机动系统的轨迹优化与控制一体化设计问题。. 通过仿真，验证了采用轨迹库的方法，可以较大地扩展吸引域；在4m/s风速（飞机巡航速度为10m/s，接近落点时的速度为2m/s至4m/s，与风速相当）的情况下，栖落机动的落点偏差不超过一个机身的长度；在风速1m/s的情况下，落点偏差在20cm左右。. 项目探讨的轨迹与控制一体化设计方法，给多约束复杂任务场合的控制系统设计提供了思路。