基于动态事件触发调度的目标状态分布式融合估计方法研究

Networked multi-sensor information fusion estimation with communication constraints has now been one of the hot research topics. The existing multi-sensor fusion estimation methods based on event-triggered scheduling only consider the communication constraint of sensor-to-fusion center forward channel, which can only reduce the total amount of communication without solving the problem of bandwidth constraint. Therefore, this project considers a networked multi-sensor distributed fusion target tracking system with feedback as the research object. In the case of sensor-to-fusion center forward channel and fusion center-to-sensor feedback channel existing communication constraints, a dynamical event-triggered scheduling strategy is proposed to meet the bandwidth constraint, and research on target state distributed fusion estimation methods based on dynamical event-triggered scheduling is carried out. The main research contents are as follows: Firstly, for the system with forward channel constraint, the multi-sensor distributed fusion estimation methods based on event-triggered scheduling are studied. Secondly, for the system with feedback channel constraint, the local sensor filtering methods based on event-triggered scheduling are discussed. Lastly, for the system with two channel constraints, the bandwidth resource scheduling strategy and event-triggered scheduling strategies of sensors and fusion center are given based on an overall performance index. The key problems should be solved such as establishment of the relationship between scheduling policy parameters and estimation performance, and establishment of the overall performance index function. This project aims to achieve some innovative research results, and lays a theoretical foundation for the methods in practical application.

具有通信约束的网络化多传感器信息融合估计是当前的研究热点。现有的基于事件触发的多传感器融合估计方法仅考虑了传感器至融合中心的前向通道约束，且只能降低总通信量而无法解决带宽受限问题。为此，本项目以网络化多传感器有反馈分布式融合目标跟踪系统为研究对象，在传感器至融合中心前向通道、融合中心至传感器反馈通道中存在通信约束情况下，提出一种动态事件触发调度策略来满足带宽约束，并开展基于动态事件触发调度的目标状态分布式融合估计方法研究。主要研究：在前向通道约束情况下基于事件触发调度的多传感器分布式融合估计方法；在反馈通道约束情况下基于事件触发调度的局部传感器滤波方法；在两通道约束情况下以综合性能指标为依据给出带宽资源调度策略以及传感器和融合中心的事件触发调度策略。重点解决建立调度策略参数和估计性能的关系、建立综合性能指标函数等关键科学问题。旨在取得一些创新性成果，为该方法在实际中的应用奠定理论基础。

本项目重点研究了网络环境下具有通信约束的网络化多传感器分布式融合估计问题。提出了一些动态调度策略来满足带宽限制的要求，如基于GSLB规则的最优调度策略、基于SRB规则的最优调度策略和基于误差阈值的事件触发调度策略。开展了基于动态事件触发调度的网络化多传感器分布式融合估计方法研究，提出了基于最小化终点估计误差的多传感器数据最优估计方法、LQR控制数据最优调度方法，并将一类系统研究推广至一般性系统，丰富了现有的研究成果。同时针对网络化系统中前向通道和反馈通道中均存在通信约束如时延及丢包的情况下，在网络化控制系统的事件触发耗散滤波、故障检测H∞滤波、最优线性指数二次高斯估计、马尔科夫跳变系统滤波等方面取得了一系列创新性成果。因此，本项目研究为网络化控制系统的应用提供了理论基础和技术支撑，具有很好的理论意义和广阔的应用前景。