Backstepping represents a promising control law for fixed-wing Unmanned Aerial Vehicles (UAVs). Its non-linearity and its adaptation capabilities guarantee adequate control performance over the whole flight envelope, even when the aircraft model is affected by parametric uncertainties. In the literature, several works apply backstepping controllers to various aspects of fixed-wing UAV flight. Unfortunately, many of them have not been implemented in a real-time controller, and only few attempt simultaneous longitudinal and lateral–directional aircraft control. In this paper, an existing backstepping approach able to control longitudinal and lateral–directional motions is adapted for the definition of a control strategy suitable for small UAV autopilots. Rapidly changing inner-loop variables are controlled with non-adaptive backstepping, while slower outer loop navigation variables are Proportional–Integral–Derivative (PID) controlled. The controller is evaluated through numerical simulations for two very diverse fixed-wing aircraft performing complex manoeuvres. The controller behaviour with model parametric uncertainties or in presence of noise is also tested. The performance results of a real-time implementation on a microcontroller are evaluated through hardware-in-the-loop simulation.

中文翻译：

---

固定翼无人机反推控制器自动驾驶仪的设计与开发

Backstepping 代表了固定翼无人机 (UAV) 的一种有前途的控制规律。即使飞机模型受到参数不确定性的影响，它的非线性和自适应能力也能保证在整个飞行包线内有足够的控制性能。在文献中，一些作品将反推控制器应用于固定翼无人机飞行的各个方面。不幸的是，其中许多还没有在实时控制器中实现，只有少数尝试同时进行纵向和横向飞机控制。在本文中，现有的能够控制纵向和横向运动的反推方法适用于定义适用于小型无人机自动驾驶仪的控制策略。快速变化的内循环变量由非自适应反步控制，而较慢的外环导航变量由比例-积分-微分 (PID) 控制。该控制器通过数值模拟对两架执行复杂机动的非常多样化的固定翼飞机进行评估。还测试了具有模型参数不确定性或存在噪声的控制器行为。通过硬件在环仿真评估微控制器上实时实现的性能结果。