CASIA OpenIR  > 毕业生  > 博士学位论文
大包线鲁棒飞行控制关键技术研究
其他题名Research on Robust Flight Control Algorithms for Large Flight Envelope
仇立伟
2010-05-31
学位类型工学博士
中文摘要飞行控制技术作为航空工业最主要的支撑技术,目前只能适应相对结构化的任务、环境变化相对不明显的要求,而距解决动态、不确定环境下的大包线自主飞行控制问题尚有很大的差距。本文应用现代控制理论的最新成果,采用自抗扰控制、鲁棒控制、反演控制等思想,以飞行姿态和轨迹为主要控制目标,针对大包线飞行控制中的难题,分别从自适应、鲁棒性能、余度思想等不同角度设计鲁棒飞行控制解决方案以确保飞行的安全、可靠。本文主要的工作和贡献如下。 (1)针对紊流风场扰动中大包线飞行器的严重非线性、快速时变、强耦合和复杂不确定性等特点,设计了面向工程应用的三通道简易自抗扰飞行控制系统,在纵向、横侧向飞行模态分别引入了神经网络对自抗扰算法中的扩张状态观测器进行合理补偿以增强对紊流风减缓能力。该控制器设计方案较传统增益调度控制策略的调参数量大为减少、鲁棒性能显著提升。 (2)提出了一种基于神经网络自适应补偿器和Backstepping算法的自适应混合鲁棒飞行控制方案来实现 高精度的飞行姿态跟踪。该方案能够适应飞行器和环境的不确定性,具有自适应变参数的能力。 (3)针对飞控系统的控制律的自适应能力受限情况,设计了基于反馈线性化和 μ 综合的无人机鲁棒混合控制器方案,反馈线性化实现飞行状态解耦, μ 综合实现鲁棒飞行品质,两者各有专攻。该方案能够适应极端的飞行环境和对象变化,具有优先保证飞控系统抗不确定性的能力。 (4)提出了一种针对飞行器协调转弯的混合非线性控制策略,该策略在飞行器的纵向通道采用了自抗扰算法实现航迹倾角的保持以达到定高的目的,横侧向通道则设计了Backstepping控制器实现滚转角保持以达到协调转弯的目的。 (5)为实现飞控系统的安全可靠性和优秀的飞行品质,提出了一种基于解析余度的飞行控制系统设计方案。该方案设计了降阶状态观测器对飞行姿态角速度的解析估计;设计了Backstepping飞行控制算法实现了高度非线性、强耦合飞行器模型的线性化和解耦;设计了扩张状态观测器来消除未建模误差和外部扰动等诸多系统不可抗因素,进而增强飞控系统的鲁棒性。 总的说来, 本文针对大包线鲁棒飞行控制算法做了有益的探索和尝试。
英文摘要Research of flight control methods is a crucial technology in aviation industry all the time. The existing flight control methods only adapt for relatively structured tasks and relatively slow-varying environment. However, since autonomous flight design in dynamic and uncertain environment hasn't been realized, there is still a long distance to go. Methods of active disturbance rejection control, robust control, backstepping control are adopted and a set of flight solution scheme is designed from adaptive, robust and redundancy perspective. The main contributions of this thesis include the following issues: (1)Focusing on severe nonlinear, coupled and time-varying plane model in large flight envelope, a simplified active disturbance rejection controller(ADRC) is designed, which is robust by introducing neural network compensation to the extended state observer of ADRC in order to enhance the ability of turbulence alleviation. Compared with traditional gain scheduling strategy, the method proposed reduces numbers of scheduling and promotes the system robustness. (2)In order to track attitude command accurately, a new hybrid controller design scheme is proposed, which consists of a backstepping controller and an adaptive neural network compensator. The scheme proposed can adapt the uncertainty of plane and environment and has ability of adaptively changing parameters. (3)The uncertainty of aircraft model far exceeds the ability of adaptive control under extreme flight environment. In order to enhance the ability of resisting uncertainty, a new hybrid controller design scheme is presented, which consists of a feedback linearization controller and a μ controller. (4)The task of coordinated turn is realized by presenting a hybrid nonlinear control strategy, which is composed of an ADRC and a backstepping controller. The ADRC of keeping flight path angle is adopted to hold altitude in longitudinal channel. The backstepping controller is designed to keeping roll angle. (5)A new scheme of analytical redundance flight control system (FCS) is proposed, which is composed of analytical redundance, core backstepping flight control algorithm and extended state observer compensator. The analytical redundancy for attitude angle rates adopts reduced order nonlinear state observer method. The backstepping flight controller realizes linearization and decoupling of the highly nonlinear and tightly coupled plane model. The extended state observer compensator is d...
关键词微分跟踪器 扩张状态观测器 神经网络 解析余度 反演控制 反馈线性化 μ 综合 Tracking Differentiator Extended State Observer Neural Network Analytical Redundance Backstepping Feedback Linearization μ Synthesis
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/6274
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
仇立伟. 大包线鲁棒飞行控制关键技术研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2010.
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_20071801462800(8505KB) 限制开放CC BY-NC-SA
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
[仇立伟]的文章
百度学术
百度学术中相似的文章
[仇立伟]的文章
必应学术
必应学术中相似的文章
[仇立伟]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。