网络化非线性系统滤波融合及控制器设计研究

Filtering fusion and controller design of networked system is one of the hot research areas. There are several problems with the existing time-delay filtering fusion methods such as constrained filtering scenes . For the respect of controller design, the majority of the exsting methods only consider certain time-delay and data packet dropout. This project focus on and solve the problem of multi-sensor nonlinear systems random delay filtering fusion and controller design for networked Lipschitz nonlinear systems. The following research will be carried out: 1) The application of filtering method based on initial state estimate and measurements weighted summation which is adopted in nonlinear filtering fusion will be carried out. The filtering strategy based on combination of prediction estimation and innovation compensation is used to handle the random delay of multi-sensor system. 2) For the multi-sensor nonlinear systems with bandwidth constraints, the strong tracking CKF filtering fusion algorithm will be given. 3) For the Lipschitz nonlinear controlled plant, considering the delay and data packet dropped out between sensor and controller also controller and actuator, the controller design method will be obtained under the condition that transition probabilities are partly unknown. This project will provide some new research ideas and solutions for filtering fusion and controller design of networked systems.

网络化系统的滤波融合及控制器设计是当前的研究热点之一。现有的网络化系统时延滤波融合方法大多存在滤波场景受限等问题，关于控制器设计大多数文献只考虑了特定的时延及丢包。本项目重点考虑和解决多传感器非线性系统随机时延滤波融合问题及网络化Lipschitz非线性系统的控制器设计问题。拟开展以下研究：1) 进行基于初始状态估计和测量值加权求和的滤波方法在非线性滤波融合中的使用研究；采用基于预测估计与新息补偿相结合的滤波策略来处理多传感器系统的随机时延问题。2) 针对带宽约束的多传感器非线性系统，给出强跟踪CKF滤波融合算法。3) 针对Lipschitz非线性被控对象，考虑传感器和控制器以及控制器和执行器之间均具有时延及丢包的情况，在转移概率矩阵部分未知的条件下，得到控制器设计方法。本项目将为网络化系统的滤波融合和控制器设计开辟一些新的研究思路和解决方案。

本项目重点解决了多传感器非线性系统随机时延滤波融合问题及网络化Lipschitz非线性系统的控制器设计问题。在滤波融合方面针对网络化系统提出了一系列滤波融合方法，如基于测量值、初始状态及标称状态，结合测量值预测和信息补偿的思路得到了一种新的扩展型Kalman滤波器。在理想通讯条件下，此滤波器和传统的Kalman滤波器相比具有相同的精度。控制器设计方面针对Lipschitz非线性被控对象，考虑传感器和控制器以及控制器和执行器之间均具有时延及丢包的情况，在转移概率矩阵部分未知的条件下，在NCS的镇定、故障检测及容错控制方面取得了一系列创新性成果。

内容获取失败，请点击重试