基于网络事件触发Markov跳变系统的异步融合估计

With the widespread application of network technology in multi-sensor information fusion estimation, in order to improve the estimation performance of the systems, the common influence of the sensor asynchronous sampling, limitations of communication energy and network bandwidth, networked uncertainties and random abrupt changes in the actual systems cannot be ignored. In the sense of linear unbiased minimum variance, this project intends to study the optimal fusion estimation problems for asynchronous Markov jump systems under the event-triggered transmission mechanism. Under different sampling schemes, the mathematical models of asynchronous Markov jump systems are established and the synchronization modeling methods of the asynchronous systems are proposed. To reduce the communication burden from sensors to estimators, the network-based event-triggered transmission strategies are proposed. The sequential fusion and the distributed fusion estimation algorithms are proposed by using the estimation theory and information fusion technology. Furthermore, the proposed algorithms are applied to solve the asynchronous fusion estimation problems for systems with time delays and packet losses. The proposed estimation algorithms will be more in line with the requirements of the actual systems for the high estimation accuracy, real-time performance and reliability of filtering estimation algorithms. The research results can be widely used in signal processing, target tracking, fault diagnosis, environmental monitoring and other fields.

随着网络技术在多传感器信息融合估计中的普遍应用，为了提高系统的估计性能，传感器异步采样、通讯能量和网络带宽的限制、网络化不确定因素以及随机突变等对实际系统的共同影响不容忽视。本项目拟在线性无偏最小方差意义下，研究事件触发传输机制下异步马尔科夫跳变系统的最优融合估计问题。建立不同采样机制下的异步马尔科夫跳变系统的数学模型，提出异步系统的同步化建模方法。为了降低传感器到估值器的通讯负担，提出基于网络的事件触发传输策略。应用估计理论和信息融合技术，提出序贯融合和分布式融合估计新算法。进而，应用所提出的算法解决带有滞后和丢包系统的异步融合估计问题。所提出的估计算法将更加符合实际系统对滤波估计算法高估计精度、实时性和可靠性的要求。研究成果可广泛应用于信号处理、目标跟踪、故障诊断和环境监控等领域。

在基于网络传输的多传感器系统中，数据传输过程中往往受能源和带宽的限制，并可能导致丢包、衰减观测和时滞等不确定性。工程实际中传感器可能异步多速率采集观测数据，并且内部参数或外部环境可能导致系统随机突变。受多种因素共同影响的多传感器状态估计问题是当前多传感器信息融合估计理论中的重点研究问题。本项目在线性无偏最小方差意义下，研究了异步多速率采样系统受能源消耗、网络化随机不确定和随机突变影响下的多传感器信息融合估计问题。通过马尔科夫链描述实际系统中的随机突变问题；采用事件触发的传输策略降低传感器到估值器和估值器到融合中心的通讯负担。利用伪测量法、迭代状态方程法、匀速增量法和状态加权法，引入狄拉克测度，通过状态增广同步化系统。应用估计理论和信息融合技术，提出了最优线性滤波器，多传感器序贯融合、分布式融合和集中式融合估计算法。研究成果可广泛应用于信号处理、目标跟踪、故障诊断和环境监控等领域。本项目研究丰富了现有的估计理论成果，拓宽了多传感器信息融合估计的思路。