芮国胜,
田文飚
海军航空大学信号与信息处理山东省重点实验室 烟台 264001
基金项目:国家自然科学基金(41606117, 41476089, 61671016)
详细信息
作者简介:董道广：男，1990年生，博士生，研究方向为贝叶斯统计学习、压缩感知和蒸发波导反演
芮国胜：男，1968年生，教授，博士生导师，主要研究方向为混沌通信系统及现代滤波理论
田文飚：男，1987年生，副教授，主要研究方向为压缩感知及蒸发波导反演
通讯作者:董道广 sikongyu@yeah.net
中图分类号:TN911.7; TP301.6计量
文章访问数:1682
HTML全文浏览量:517
PDF下载量:38
被引次数:0
出版历程
收稿日期:2019-07-25
修回日期:2020-03-26
网络出版日期:2020-04-24
刊出日期:2020-07-23
Research on the Dynamic Sparse Bayesian Recovery of Multi-task Observed Streaming Signals in Time Domain
Daoguang DONG,,Guosheng RUI,
Wenbiao TIAN
Signal and Information Processing Key Laboratory in Shandong, Navy Aviation University, Yantai 264001, China
Funds:The National Natural Science of China (41606117, 41476089, 61671016)
摘要
摘要:为了解决多任务观测条件下时域流信号动态重构面临的块效应问题，该文基于重叠正交变换(LOT)和稀疏贝叶斯学习的贪婪重构框架先后提出了一种流信号多任务稀疏贝叶斯学习算法及其鲁棒增强型的改进算法，前者将LOT时域滑窗推广到多任务条件下，通过贝叶斯概率建模将未知的噪声精度的估计任务从信号重构中解耦并省略，后者进一步引入了重构不确定性的度量，提高了算法的鲁棒性和抑制误差积累的能力。基于浮标实测数据的实验结果表明，相比多任务重构领域代表性较强的时间多稀疏贝叶斯学习(TMSBL)和多任务压缩感知(MT-CS)算法，本文算法在不同信噪比、观测数目和任务数目条件下具有显著更高的重构精度、成功率和效率。
关键词:信号处理/
流信号/
多任务/
稀疏贝叶斯/
块效应
Abstract:To eliminate the blocking effects in the dynamic recovery of the streaming signals observed from multiple tasks in time domain, a streaming multi-task sparse Bayesian learning based algorithm and its robust enhanced version are proposed in this paper, where the former extends Lapped Orthogonal Transform (LOT) sliding window in time domain to multi-task condition, and decouples the estimation of unknown noise accuracy from signal reconstruction by Bayesian probability modeling and omits it, the latter further introduces the measurement of reconstructed uncertainty, which improves the robustness of the algorithm and the ability to suppress the accumulation of errors. Experimental results based on measured meteorological data shows that the proposed algorithms have significantly higher reconstruction accuracy, success rate and running speed than the representative algorithms in the field of compressed sensing from multiple measurement vectors, namely, the Temporal Multiple Sparse Bayesian Learning (TMSBL) algorithm and the Multi-Task-Compressed Sensing (MT-CS) algorithm, under different conditions of Signal-to-Noise Ratios, number of observations and tasks.
Key words:Signal processing/
Streaming signals/
Multi-task/
Sparse Bayesian/
Blocking effects
PDF全文下载地址:
https://jeit.ac.cn/article/exportPdf?id=89cf7b72-9d12-4815-bba3-e6292ed02b45
