CASIA OpenIR  > 毕业生  > 博士学位论文
无人机鲁棒非线性飞行控制技术研究
其他题名Robust Nonlinear Flight Control for Unmanned Aerial Vehicles
袁如意
学位类型工学博士
导师郁文生 ; 易建强
2011-05-27
学位授予单位中国科学院研究生院
学位授予地点中国科学院自动化研究所
学位专业控制理论与控制工程
关键词飞行控制 非线性控制 鲁棒控制 反馈线性化 反馈线性化 H∞控制 Backstepping算法 Flight Control Flight Control Robust Control Feedback Linearization Saturation Control H∞control Backstepping
摘要飞行控制系统是无人机系统的核心部分, 无人机各项性能在很大程度上取决于飞行控制系统的性能. 无人机是一个高度非线性、快时变、强耦合的复杂系统, 其精确模型很难获得, 且运行环境中存在着各种干扰, 如传感器的量测噪声与大气中的紊流等. 如何克服不确定性与扰动、设计高性能的飞行控制律是无人机飞控系统设计中所面临的首要问题和难题. 传统线性方法已难以满足现代飞控设计的要求, 以无人机非线性模型为基础设计飞行控制律已成为当前研究热点. 本文以无人机的非线性飞行控制为背景,深入研究无人机的非线性鲁棒飞行控制技术, 主要完成以下工作: 1. 对无人机模型中的非线性、不确定性及紊流对无人机的影响进行了分析. 2. 提出一种基于微分几何反馈线性化与扩张状态观测器的无人机鲁棒姿态控制算法. 通过反馈线性化使姿态系统成为已解耦但存在不确定性及外界扰动的线性对象, 利用扩张状态观测器估计不确定性及扰动并在控制律中实现补偿, 实现鲁棒的、精确的姿态控制. 在此基础上设计了鲁棒的协调转弯控制器. 仿真验证了算法的有效性. 3. 提出一种基于Backstepping算法和非线性H∞<上标!>控制的无人机机动飞行控制算法. 利用Backstepping算法设计的灵活性构造控制律与代价函数, 使得闭环系统及代价函数满足Hamilton-Jacobi-Isaacs(HJI)方程从而具有渐近跟踪性能和干扰抑制能力, 实现鲁棒的机动飞行控制, 仿真验证了算法的有效性. 4. 对一类未知非线性系统, 提出一种基于自适应神经网络的、考虑控制输入约束的H∞<上标!>跟踪控制算法, 并将其与动态逆算法结合设计了鲁棒的飞行控制律. 该算法能将飞行器的作动器 动态融合到控制器的设计过程中, 使得根据控制器得到的控制量满足作动器速度与幅度约束, 且可以通过参数来调整跟踪效果与扰动抑制效果. 仿真验证了算法的有效性. 最后对全文的工作进行了总结, 并指出鲁棒非线性飞行控制中存在的问题和一些需要进一 步研究的工作.
其他摘要Unmanned Aerial Vehicles (UAVs) flight control system is the core of the UAV system, the performance of UAV depends largely on the UAV flight control system’s performance. UAV is a high nonlinear, fast time-varying, and strong coupled system, whose accurate mathematical model is difficult to be obtained and operating environment is full of disturbance such as sensor noise and atmospheric turbulence. How to overcome the uncertainty and undesired disturbance to obtain robust and high-performance flight control laws is the most important issue in the flight control system design for UAV. Traditional linear design methods have been difficult to meet the requirements of modern flight control. Nonlinear model-based UAV control law design has already been focused by more and more researchers. In this dissertation, based on nonlinear flight control law design for UAVs, robust nonlinear flight control problem is deeply researched and the following progresses have been made. 1. The effects of nonlinearity, uncertainty in UAVs model and atmospheric turbulence to UAVs are analyzed. 2. A robust attitude control algorithm is proposed based on Extended State Observer and feedback linearization via differential geometry. The attitude system is decoupled into linear plants with the existences of uncertainty and extern disturbance. In order to achieve robust and accurate attitude control, Extended State Observers are used to estimate uncertainty and extern disturbance so as to compensate them in control law design. A robust coordinated turn controller is designed based on the proposed algorithm. The validity and advantages of the algorithm and the coordinated turning controller are demonstrated by numerical simulations. 3. A maneuver control algorithm for UAVs is proposed based on Backstepping algorithm and nonlinear H∞control theory. The control laws and cost functions are constructed by the designing flexibility of Backstepping algorithm so that the closed-loop systems and the cost functions satisfy Hamilton-Jacobi-Isaacs(HJI) equations, thus guaranteeing the closed-loop systems asymptotic tracking ability and disturbance attenuation ability. Simulation results show the validity of the algorithm. 4. For a class of unknown nonlinear system, an adaptive neural network based H∞tracking control algorithm which can handle the constraints on control inputs is proposed. Then a robust flight control algorithm is designed using neural network based H∞tracking co...
馆藏号XWLW1602
其他标识符200818014628028
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/6348
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
袁如意. 无人机鲁棒非线性飞行控制技术研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2011.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
CASIA_20081801462802(6307KB) 暂不开放CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[袁如意]的文章
百度学术
百度学术中相似的文章
[袁如意]的文章
必应学术
必应学术中相似的文章
[袁如意]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。