基于FNN的随机非线性系统控制器设计与分析

发布时间：2018-10-25 17:53

【摘要】：实际工程系统往往表现为非线性系统,而且系统中不可避免地存在各种不确定性和外部随机扰动。这些不确定性和随机扰动会对系统的稳定性和性能产生影响。因此,随机非线性系统的控制问题研究具有重要的理论意义。在工程实践中,如何克服不确定性和随机扰动对飞行控制系统稳定性的影响进而保证飞行安全,是具有实际应用价值的课题。因此,本论文主要针对随机非线性不确定系统进行镇定控制器的分析和设计,并将理论研究成果应用于飞行系统的稳定性控制中。主要研究内容如下:首先对国内外随机非线性系统控制问题研究的发展现状进行介绍总结,分析目前研究热点与需解决的问题;在随机Lyapunov稳定性理论框架下,将Backstepping技术与模糊神经网络方法相结合,通过构建四层模糊神经网络,使得输出权值可自适应调整,从而设计出使得一类纯反馈随机非线性不确定系统的状态依概率有界的自适应控制器,设计方法有效地减少了可调参数的数目;针对Backstepping的不足,为改进控制器设计,引入动态面控制方法,通过一阶低通滤波器的应用避免参数膨胀,并简化了实际控制器结构,降低计算量,设计出使得闭环系统半全局一致最终有界的自适应控制器;此外,通过减小滤波时间来加快系统状态收敛且减小滤波误差;将所获得的理论成果应用于含随机扰动的高超速飞行器纵向模型控制中,设计出自适应模糊神经网络动态面控制器,以保证闭环系统信号是半全局一致最终有界的,仿真结果验证了该方法的有效性;最后,对全文进行总结与展望。
[Abstract]:Practical engineering systems are often nonlinear systems, and there are inevitably various uncertainties and external random disturbances in the systems. These uncertainties and random disturbances will affect the stability and performance of the system. Therefore, it is of great theoretical significance to study the control problem of stochastic nonlinear systems. In engineering practice, how to overcome the influence of uncertainty and random disturbance on the stability of flight control system and ensure flight safety is a subject of practical application value. Therefore, the stabilization controller of stochastic nonlinear uncertain systems is analyzed and designed in this paper, and the theoretical research results are applied to the stability control of flight systems. The main research contents are as follows: firstly, the development of stochastic nonlinear system control problems at home and abroad is introduced and summarized, and the current research hotspots and problems to be solved are analyzed. Combining Backstepping technology with fuzzy neural network method, the output weight can be adjusted adaptively by constructing a four-layer fuzzy neural network. Thus, an adaptive controller is designed, which makes the state of a class of pure feedback stochastic nonlinear uncertain systems bounded by probability. The design method effectively reduces the number of adjustable parameters. The dynamic surface control method is introduced to avoid the parameter expansion through the application of the first-order low-pass filter. The structure of the controller is simplified, the computation is reduced, and an adaptive controller is designed to make the closed-loop system semi-global uniform and ultimately bounded. The filtering time is reduced to accelerate the state convergence of the system and the filtering error is reduced. The obtained theoretical results are applied to the longitudinal model control of high speed vehicles with random disturbances, and an adaptive fuzzy neural network dynamic surface controller is designed. In order to ensure that the closed-loop system signal is semi-global uniform and ultimately bounded, the simulation results verify the effectiveness of the method. Finally, the paper is summarized and prospected.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273

【参考文献】