【作者】 于洋；

【导师】 袁建平；

【作者基本信息】 西北工业大学 ， 飞行器设计， 2022， 博士

【摘要】 随着空间探索、开发与应用能力的不断提高,面向各种任务需求的、结构日趋复杂的航天器相继进入太空。然而航天器在执行空间操控任务过程中不可避免的受到外部环境干扰力矩、未建模动态、过程/测量噪声、执行器饱和约束等物理层中的不确定性,此外还受到通信带宽约束、恶意的网络攻击等网络层中的不确定性因素。如何保证航天器在多源不确定性的复杂空间环境中更加持久、稳定、安全、高质量地在轨运行,已成为目前航天技术领域亟待解决的关键问题。本文针对航天器姿态跟踪控制问题,在综合考虑系统物理层与网络层中的多源不确定性的情况下,提出了一系列形式的扰动/攻击观测器,并基于不变性原理,设计前馈与反馈相结合的复合抗干扰及安全控制器,旨在增强航天器在姿态跟踪过程中的鲁棒性、安全性及可靠性。本文的主要研究内容和研究成果如下:(1)针对含有执行器饱和与未知扰动的航天器相对姿态采样输出系统,开展了一种基于混杂型扩张状态观测器的复合抗干扰控制研究。为了减轻采样输出诱导出的“栅栏效应”,利用采样间预估器与扩张状态观测器设计了一种混杂型扩张状态观测器,以采样输出数据为输入对系统状态及未知项进行实时估计;采用“前馈+反馈”形式设计一种确保执行机构安全运行的复合抗干扰控制器。相比于基于零阶保持器的控制策略,该方法对系统瞬态性能与稳态性能均起到明显的提升作用。(2)针对含有通信带宽约束及传输野值的航天器相对姿态非线性系统,开展了一种基于强制型扩张状态观测器的复合抗干扰控制研究。采用静态事件触发机制减少测量数据的传输率;基于饱和函数幅值限制约束与混杂型扩张状态观测器设计了一种强制型扩张状态观测器,削弱测量数据中野值的影响,同时对系统中的状态及未知项进行实时估计,并设计复合抗干扰控制器对扰动进行补偿。该控制策略利用有限的且含有野值的测量数据依然能够实现较高精度的状态估计与补偿,显著增强了观测器对测量数据的鲁棒性。(3)针对含有多源不确定的航天器相对姿态精细抗干扰控制问题,开展了一种基于复合干扰观测器的反步控制策略研究。采用反步法设计虚拟控制器将系统中的非匹配扰动当作匹配扰动进行处理;基于神经网络、干扰观测器及扩张状态观测器设计一种新型的复合干扰观测器,对不同属性的多源匹配扰动做到“知己知彼并对症下药”,同时对多源不确定项实施精细估计,并设计“多前馈+反馈”形式的复合抗干扰控制器。该控制策略的提出降低了将多源不确定性当作“总和扰动”处理的保守性,并且提高了系统的估计与控制精度。(4)针对传感器端网络遭受欺骗攻击的航天器相对姿态非线性系统,开展了一种基于自适应扩张状态观测器的复合抗干扰控制策略研究。通过设计基于动/静态事件触发机制的自适应扩张状态观测器,利用有限且受恶意攻击的测量信息,对系统中的状态及未知非线性项实时估计,并设计复合抗干扰控制器对物理层的扰动实施前馈补偿。该控制策略既能够抑制网络层的欺骗攻击又能抵消未知非线性项对系统的影响。(5)针对通信网络遭受隐蔽型攻击的航天器相对姿态线性系统,开展了一种基于自适应观测器的复合抗干扰及抗攻击控制策略研究。采用卡尔曼滤波技术与扩张状态观测器相结合的思想,分别利用传感器攻击后的信息与真测量实信息设计自适应观测器克服系统噪声的同时对系统中的状态、外界干扰及攻击信号进行实时估计;并设计“多前馈+反馈”形式的复合抗干扰及抗攻击控制器。该控制策略的提出抑制了噪声对系统控制精度的影响,并提高了系统抗攻击性能,为确保航天器主动抗干扰及抗攻击策略的实施奠定基础。本文提出的一系列基于扩张状态观测器的多源不确定性的估计与补偿算法,不仅适用于航天器姿态控制系统,也为其他工程应用的非线性网络化系统的主动抗干扰及抗攻击控制提供了新的思路。更多还原

【Abstract】 With the continuous improvement of space exploration,development and application capabilities,spacecraft with increasingly complex structures have entered space in succession to meet the needs of various tasks.However,spacecraft are inevitably subjected to uncertainties in the physical layer,such as external environment interference torque,unmodeled dynamics,process/measurement noise,actuator saturation constraints,as well as the uncertainties in the network layer such as communication bandwidth constraints and malicious network attacks.How to ensure that spacecraft can operate in orbit more permanently,stably,safely and with high quality in the complex space environment with multi-source uncertainty has become a key problem to be solved urgently in the field of space technology.In this paper,a series of disturbance/attack observers are proposed for spacecraft attitude tracking control,considering multi-source uncertainties in the system physical layer and network layer.Based on the invariance principle,some composite anti-disturbance and safety controllers combining feed-forward and feedback are designed to enhance the robustness,safety and reliability of spacecraft in the attitude tracking process.The main contents and achievements are summarized as follows:(1)A composite anti-disturbance control method based on hybrid extended state observer is developed for spacecraft relative attitude sampling-data system with actuator saturation and unknown disturbances.In order to reduce the "fence effect" induced by the sampling output,a hybrid extended state observer is designed by using the inter-sample predictor and extended state observer,which used the sampling output data as input to estimate the system states and unknown terms in real time.A composite anti-disturbance controller is proposed to ensure the safe operation of the actuator by adopting the form of "feedforward + feedback".Compared with the control strategy based on zero order holder,the proposed method significantly improves the transient and steady-state performance of the system.(2)For the nonlinear spacecraft relative attitude system with communication bandwidth constraint and transmission outliers,a composite anti-disturbance control based on stubborn extended state observer is developed.A static event-triggered mechanism is introduced to reduce the transfer rate of the measured data.Based on the characteristics of saturation function and hybrid extended state observer,a stubborn extended state observer in proposed to weaken the influence of outliers,and to estimate the states and unknowns in the system in real time.A composite anti-disturbance controller is designed.The proposed control strategy can still achieve high precision state estimation and compensation by using the limited measured data with outliers,which significantly enhances the robustness of the observer to the measured data.(3)In order to solve the problem of precise anti-disturbance control of spacecraft relative attitude with multi-source uncertainties,a backstepping control strategy based on composite disturbance observer is developed.The virtual controller is designed by backstepping method to overcome the unmatched disturbance treated as matched disturbance.Based on neural network,disturbance observer and extended state observer,a new type of composite disturbance observer is designed to "know oneself and the enemy and take appropriate medicine" for multi-source matching disturbances with different attributes.At the same time,refined estimation of multi-source uncertainties is implemented.A composite antidisturbance controller with the form of "multi-feedforward + feedback" is designed.The proposed control strategy reduces the conservatism of treating multi-source uncertainty as "total disturbances" and improves the estimation and control accuracy of the system.(4)A composite anti-disturbance control strategy based on adaptive extended state observer is developed for the nonlinear spacecraft relative attitude system with deception attack on the sensor network.To be specific,an adaptive extended state observer based on dynamic/static event-triggered mechanism is introduced to estimate the states and unknown nonlinear terms in real time using the limited and malicious attack measurement information.Then a composite anti-disturbance controller is designed based on the estimations.The control strategy can not only suppress the deception attack at the network layer,but also overcome the influence of unknown nonlinear terms on the system.(5)A composite anti-disturbance and anti-attack control strategy based on adaptive observer is developed for the spacecraft relative attitude linear system with covert attack in the communication network.Using Kalman Filter technology and extended state observer,two adaptive observers are proposed by using the sensor attacked information and real information,respectively,to overcome the system noise and estimate the state,external disturbance and attack signal in real time.And a composite anti-disturbance and anti-attack controller with the form of "multi-feedforward + feedback" is designed.The proposed control strategy can suppress the influence of noise on the control precision and improve the anti-attack performance of the system.It lays a foundation for the implementation of spacecraft active anti-disturbance and attacks strategies.To conclude,a series of multi-source uncertainty estimation and compensation algorithms based on extended state observer are proposed,which are not only suitable for spacecraft attitude control systems,but also provide a new idea for active anti-disturbance and anti-attack control of nonlinear networked systems in other engineering applications.更多还原