自适应两阶段非线性容积Kalman滤波融合方法研究

项目名称： 自适应两阶段非线性容积Kalman滤波融合方法研究

项目编号： No.61503213

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 张露

作者单位： 衢州学院

项目金额： 21万元

中文摘要： 多传感器移动平台防空系统信息融合作为打赢信息化战争的先决条件和决定因素，有着巨大的发展潜力，也面临着严峻的挑战。平台中多样的偏差源和强耦合的系统偏差使得现有信息融合方法已经难以满足防空系统日益增长的精准跟踪需求。因此，本项目以多传感器移动平台中复杂非线性目标跟踪系统为对象，以两阶段滤波思想为主体，以高性能的容积Kalman滤波为基础，以变分贝叶斯、Sage-Husa方法、强跟踪滤波理论和带有反馈机制的性能评估方法为技术手段，开展自适应两阶段非线性容积Kalman融合方法研究，重点解决所遇到的关键性科学问题：1）基于状态与系统偏差联合估计的非线性系统偏差估计方法；2）基于渐消因子和模型参数自适应估计方法的两类自适应两阶段非线性容积Kalman估计融合方法；3）基于分步估计结果可用性的反馈式自适应两阶段容积Kalman滤波融合方法。最后，基于仿真和实际数据开展相应算法和估计融合方法验证与测试。

中文关键词： 多传感器移动平台；两阶段容积Kalman滤波；系统偏差；估计融合；自适应方法

英文摘要： Information fusion of multi-sensor mobile platform air defense system as a prerequisites and determinants for information warfare，not only has the huge development potential, but also faces severe challenges. For the diversities of bias source and strong coupling of system bias in platform，it is difficult to satisfy the increasing demand for air defense system with accurate tracking. Aiming at this complex nonlinear target tracking system in multi-sensor mobile platform, we develop the study on adaptive nonlinear two-stage cubature Kalman fusion method based on high-performance cubature Kalman filtering and some technologies such as Variational Bayesian method, Sage-Husa method, strong tracking filtering theory and Performance evaluation method with feedback mechanism. The following key problems should be solved: the nonlinear system bias estimations based on Joint Estimation of state and system bias, two classes adaptive nonlinear two-stage cubature Kalman filtering fusion method on the basis of fading factor and model parameter adaptive estimation method, and feedback Type adaptive two-stage cubature Kalman filtering fusion method using results usability of step by step estimation. At last, the validation of the related algorithms and estimation fusion methods should be developed on basis of simulated data and practical data.

英文关键词： Multi-sensor mobile platform;Two-stage Cubature Kalman filtering;System bias;Estimation fusion;Adaptive method