双跳变目标跟踪系统的分布式估计

项目名称： 双跳变目标跟踪系统的分布式估计

项目编号： No.61203044

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 自动化学科

项目作者： 李文玲

作者单位： 北京航空航天大学

项目金额： 25万元

中文摘要： 本项目以无线传感器网络中的移动终端定位为背景，针对移动终端的机动和量测信息在视线/非视线之间的切换建模为不同Markov跳变过程的特点，研究双跳变目标跟踪系统的分布式估计问题。为克服通过交互量测信息的传统单模型分布式实现方法给多模型估计带来执行滤波器指数增长的困难，本项目在Bayes估计框架下，拟采用定义等价信息量的方法，确定各传感器与相邻传感器之间交互的信息，系统提出针对多模型的分布式估计方法并进行性能分析。然后，考虑到网络带宽的限制，研究量化通信约束下分布式多模型估计问题，并以多移动机器人跟踪为应用对象进行仿真与实验验证。通过本项目的研究不仅能够提供新的分布式估计和多模型估计方法，而且为解决复杂环境中分布式多传感器多机动目标跟踪问题提供坚实的理论指导和有效的技术支撑，具有重要的科学意义与应用价值。

中文关键词： 随机跳变系统；分布式估计；多机动目标跟踪；；

英文摘要： This project mainly studies the problem of distributed estimation for target tracking systems with two jumping parameters. This is inspired by the mobile terminal location in the wireless sensor networks, in which the features of the mobile terminal maneuvers and the switching of the line-of-sight/non-line-of- sight measurements are modeled as different jumping Markov processes.To overcome the difficulty of exponentially growing filters in the multiple model estimation by communicating measurements between neighboring sensors, appropriate equivalent quantities are defined for the measurements in the framework of Bayes estimation so that the distributed multiple model estimator and its performance analysis can be developed. Furthermore,the problem of distributed estimation with quantized communication is investigated due to bandwidth limitations in the sensor networks. The theoretical results will be verified by experiments involving multiple mobile robots tracking. This project is of great significance in theoretical and practical applications since it not only develops novel methods for distributed estimation and multiple model estimation but also provides solid theoretical guidance and effective technical support for distributed multi-sensor multiple maneuvering targets tracking in complex scenarios.

英文关键词： Stochastic jump system；Distributed estimation；Multiple maneuvering target tracking；；