小型无人直升机姿态非线性鲁棒控制设计

Nonlinear robust attitude control for a miniature unmanned helicopter

中文关键词  无人直升机  参数辨识  神经网络  滑模控制

英文关键词  unmanned helicopter  parameter identification  neural networks  sliding-mode control

中文摘要

      本文针对小型无人直升机的姿态控制问题, 通过系统参数辨识, 获得了较为准确的无人直升机姿态动力学模型. 并根据无人直升机的动态特性, 设计了基于神经网络前馈与滑模控制的非线性鲁棒姿态控制律, 该控制律对直升机模型的先验知识要求较低. 利用基于Lyapunov的分析方法证明, 设计的控制律能够实现对无人直升机姿态角的半全局指数收敛镇定控制, 并能确保闭环系统的稳定性. 基于姿态飞行控制实验平台的实时飞行控制实验结果表明, 提出的控制设计取得了很好的姿态控制效果, 并对系统不确定性和外界风扰动具有较好的鲁棒性.

英文摘要

      The attitude dynamic model for a miniature unmanned helicopter is obtained via system identification. To deal with the modeling uncertainties and existing external disturbances, we combine a neural-network-based feedforward term with the sliding-mode control method to formulate a new nonlinear robust attitude controller for the unmanned helicopter. The proposed controller requires very little model knowledge for the helicopter dynamics, and is with good robustness against model uncertainties and external disturbances. Lyapunov stability analysis is employed to show that the proposed control design can achieve semi-global exponential tracking result of the helicopter attitude angles while keeping all the closed-loop system signals bounded. To validate the performance of the proposed control design, we implement a real-time experiment on a self-built test-bed. The experiment results show that the controller developed in this paper achieves good attitude control performance under model uncertainties and external wind disturbances.