【作者】 冯伟；

【导师】 龚振邦； 罗均；

【作者基本信息】 上海大学 ， 机械电子工程， 2009， 博士

【摘要】 超小型无人直升机具有成本低、体积小、携带方便、起飞着陆场地小、可以滞空飞行等特点,使其在军用和民用领域都有着广阔的应用前景。超小型无人直升机是一个多变量、非线性、强耦合的动力学系统,稳定性差,容易受到风等外力的干扰。飞行控制系统是超小型无人直升机实现自主飞行、提升应用价值的最为核心的一项技术,而飞行姿态控制又是其中的关键。鲁棒控制对多变量控制系统和存在干扰情况的控制系统具有优势,因此对超小型无人直升机飞行姿态进行鲁棒控制的研究,具有重要的理论意义和应用价值。本论文在分析了国内外对超小型无人直升机飞控系统研究现状的基础上,归纳了超小型无人直升机飞控系统的结构和特点。根据研究背景的需要,提出了进行超小型无人直升机飞控系统的研究目标,规划了为实现研究目标而需要进行的研究内容。本论文研究首先从实际需求出发,基于模块化的思想提出了超小型无人直升机姿态控制的总体方案,并对超小型无人直升机系统的传感器、控制器和通讯系统等进行了选型。超小型无人直升机飞行运动数学模型是设计控制律的基础,飞行品质的好坏也严重依赖于数学模型建立的精确与否。在建模方面,本文建立了超小型无人直升机的机体刚体动力学方程和空气动力学模型,同时对主桨和贝尔-希勒翼进行了研究,在结构分析的基础上,基于小干扰情况下的合理假设,通过把非线性模型在悬停点附近线性化,得到用物理参数描述的线性模型,从而确定了具有物理意义的超小型无人直升机姿态模型的结构。对模型中存在的未知参数,本文采用指定状态变量的子空间辨识方法,在开环系统的子空间辨识算法的基础上,结合主元分析法,通过正交子空间投影的闭环子空间辨识方法,进行系统辨识,从而获得系统的主要结构参数,并估算出这些参数的误差范围。然后将所得模型的预测数据与飞行数据进行了比较。为了子空间系统辨识的需要,本文将超小型无人直升机姿态模型进行了离散化处理,得到了离散化的模型结构。对于试验数据,进行了野值剔除、插值、滤波、平滑等数据处理,以使得数据更精确可靠。在控制系统设计中采用的模型与实际控制对象存在着一定的差异,即存在着模型不确定性。同时,控制系统的运行也受到周围环境和有关条件的制约。这就要求超小型无人直升机飞控系统具有良好的处理模型不确定性及外界扰动的能力,即具有鲁棒性。针对通过辨识得到的系统模型,对于小干扰的情况,本文推导了离散系统不确定性具有块对角结构的鲁棒保性能控制律设计方法,并以超小型无人直升机为对象,设计了一个鲁棒保性能控制器,使得闭环系统对所有允许的不确定性渐近稳定,并且闭环性能指标不超过某个确定的上界。对于较大干扰的情况,本文推导了离散系统不确定性具有块对角结构的鲁棒方差控制律设计方法和控制能量具有上界的鲁棒方差控制律设计方法,并以超小型无人直升机为对象,设计了一个控制能量具有上界的鲁棒方差控制器,使得系统在有外部干扰的情况下具有良好的稳态性能和良好的瞬态性能,并且其输入能量不大于确定的上界。然后,分别对保性能控制系统和鲁棒方差控制系统的最坏情况进行了定义,并对最坏情况下的仿真进行了理论推导和仿真。在上述研究的基础上,本论文对超小型无人直升机飞控系统硬件、飞控系统软件和地面监控站进行了研究和设计。系统硬件方面,在任务分解和模块化设计思想的基础上,主要设计了传感器模块、模式切换与安全保障模块、主控模块;飞控系统软件方面主要着眼于任务分解、任务调度和通讯协议方面的研究;在地面监控站方面,在分析其功能需求的基础上,着重于采用组件化的编程方法。最后,通过超小型无人直升机系统完成姿态控制飞行实验(直升机系统稳定性最差的飞行状态),验证了所设计的超小型无人直升机飞行姿态鲁棒控制系统的性能。通过分析试验数据验证了超小型无人直升机系统姿态控制算法的可行性。更多还原

【Abstract】 Subminiature Unmanned Helicopter(SUH) has the advantage of the low cost、the small volume、the convenience for transportation、the small land for flying-off or landfall and hang in the small space. So it has the widely purpose in the military and civil fields. The dynamics of a SUH has the character of multivariable,nonlinear and strong coupling. So the SUH can be interfered with wind easily and hard to keep stable. The Flight Control System (FCS) is the pivotal part to realize autonomous flight and promote the practicability of the SUH and the attitude control is the key in the FCS. Robust control theory is one of the popular areas in the field of control engineering to deal with multivariable system and to improve the anti-noise ability. So the research of the robust attitude control for the SUH is very valuable.Foremost, clued by some overseas successful SUHs, the state-of-the-art of SUHs was studied, the frameworks and the characters of the FCSs were concluded and analyzed, and then the goal and the contents of the research were decided.From the view of the actual requirement, the whole scheme of system was designed on the thought of modularization, and the sensors、the microchip and data linker of FCS were chosen.The flight motion model of the SUH is the precondition of flight control laws design and the flight quality is also depending on the model. For the model of the SUH, the geostatic equation and aerodynamics equation was set up. At the same time, the main blade and the bell-hiller flybar were studied. Based on the above research and reasonable assumption,through the linearization of the nonlinear dynamics, the linear time-invariant attitude dynamics with physical parameters was obtained.To the unknown parameters of the above attitude model, a subspace based state space system identification method based on principal component analysis was carried out,which can identify the state space model of a plant with white noise and controller through orthogonal projection. And the range of the error of the parameters was also been estimated. Then the valid of the identified attitude model was testified. In order to apply the identification,the attitude model was been discretized firstly. Then the data processing of the flight data which to deal with the occasional bursts and outliers, missing data etc was processed to make sure that the data would be more reliable and correct.In general, the structure uncertainty or parameter uncertainty exist in almost systems for the tolerance between execution unit and control unit, and also the system was interfered with outer circumstances inevitably. So controller and analysis outcome of systems with precise mathematical models are always dissatisfied with project requirement. Aimed at the attitude model gotten form the identification, an approach for guaranteed cost control was theoretical derived for the system with the diagonally uncertainty for small noise circumstance,and the cost no more than a supremum. An approach for robust covariance control and another for which has limited input power was theoretical derived for the system with the diagonally uncertainty for noise circumstance which made the system has good steady performance and transient property. Then the worst instances were been defined for the guaranteed cost control and robust covariance control, and the simulation was theoretical derived and been done.On the basis of the above research, the hardware and software of the FCS and the Ground Control Station (GCS) was studied and developed then. With respect to the hardware of FCS, sensors module, remotely control/automatic control switch circuits and the main module were designed based on task sorting and the thought of modularization. With respect to the software of FCS, the task sorting, the task scheduling module and the communication protocols were designed. With respect to the GCS, founded on the analysis of the functional requirements of the GCS, the component programming method based on component object model was considered.Finally, the flight experiments are carried through by the autonomously hanging flight which is the worst flight state based on the FCS and GCS of this paper, the experiment results validate that the attitude is stable in flight under the action of FCS by analyzing the flight data.更多还原