基于ADRC迭代学习控制的四旋翼无人机姿态控制

针对农用无人机超低空表型遥感和喷药精准悬停易受地效扰动问题，提出了一种自适应ADRC姿态控制器。首先设计了基于ADRC的姿态控制器，结合四旋翼无人机平台在0.9~1.1、1.1~1.3、1.4~1.6、2.0~2.4、2.5~2.9、3.3~3.6 m/s侧向水平风、0.9~1.1 m/s （11°）、1.1~1.3 m/s （13°）、1.4~1.6 m/s （18°）、1.8~2.0 m/s （18°）、2.1~2.5 m/s （18°）前俯向风和侧俯向风下进行干扰的预测和控制量的补偿实验。实验结果显示使用ADRC姿态控制器后无人机抗风性能有较大提升。然而在存在初始误差时，ADRC固定带宽无法满足要求，进一步设计了自适应ADRC姿态控制器（ILC-ADRC）。通过迭代学习控制在线优化自抗扰控制器带宽，实现了不同增益观测器的自适应整定。实验结合四旋翼无人机平台分别进行了机头实际方向与期望方向偏离55°、90°、180°，水平风速1.1~1.3、1.4~1.6、2.0~2.4、2.5~2.9 m/s下使用ADRC和ILC-ADRC的对比。实验结果显示采用ILC-ADRC姿态控制器，在150次控制周期内，偏航角误差均在-15°~15°之间，满足四旋翼无人机偏航角控制精度要求，同时调节时间分别缩短了40%，16.67%，12.5%，53.33%，10.34%，13.95%，27.27%，58.66%，11.86%。

Active disturbance rejection control (ADRC)of UAV attitude based on iterative learning control

An adaptive Active Disturbance Rejection Control (ADRC) attitude controller is proposed to solve the problem that the ultralow altitude phenotype remote sensing and precise hovering during spraying of agricultural UAVs are easily disturbed by ground effects. First, an attitude controller based on ADRC is designed, and an interference prediction and control amount compensation experiment is performed under lateral horizontal wind of 0.9-1.1, 1.1-1.3, 1.4-1.6, 2.0-2.4, 2.5-2.9, 3.3-3.6 m/s, and forward pitch and side pitch wind of 0.9-1.1m/s (11 °), 1.1-1.3 m/s (13 °), 1.4-1.6 m/s (18 °), 1.8-2.0 m/s (18 °), 2.1-2.5 m/s (18 °). The experimental results show that the wind resistance of the UAV has been significantly improved by the ADRC attitude controller. However, the ADRC fixed bandwidth cannot meet the requirements when an initial error exists. Therefore, an adaptive ADRC attitude controller (ILC-ADRC) is further designed to optimize the bandwidth of the ADRC controller online to achieve adaptive tuning of different gain observers. Experiments are conducted to deviate the actual direction of the head from the desired direction by 55 °, 90 °, 180 °, with the horizontal wind speeds of 1.1-1.3, 1.4-1.6, 2.0-2.4, 2.5-2.9 m/s. The results show that with the ILC-ADRC attitude controller, the yaw angle error is within -15 ° -15 ° within 150 control cycles, satisfying the control accuracy of the yaw angle of the four-rotor UAV, and the stability time is shortened by 40%, 16.67%, 12.5%, 53.33%, 10.34%, 13.95%, 27.27%, 58.66%, 11.86%, respectively.