[1]周林,刘先省,方拥军,等.贝叶斯框架下基于凸优化的系统偏差估计方法[J].探测与控制学报,2019,41(04):92.[doi:.]
 ZHOU Lin,LIU Xianxing,FANG Yongjun,et al.Systematic Biases Estimation Based on Convex Optimization under Bayesian Framework[J].,2019,41(04):92.[doi:.]
点击复制

贝叶斯框架下基于凸优化的系统偏差估计方法()
分享到:

《探测与控制学报》[ISSN:1008-1194/CN:61-1316/TJ]

卷:
41
期数:
2019年04
页码:
92
栏目:
出版日期:
2019-08-26

文章信息/Info

Title:
Systematic Biases Estimation Based on Convex Optimization under Bayesian Framework
文章编号:
1008-1194(2019)04-0092-06
作者:
周林刘先省方拥军金勇
河南大学计算机与信息工程学院,河南 开封 475004
Author(s):
ZHOU LinLIU XianxingFANG Yongjun JIN Yong
Computer and Information Engineering College, Henan University, Kaifeng, 475004, China
关键词:
系统偏差估计配准凸优化贝叶斯框架
Keywords:
systematic biases estimation convex optimization Bayesian framework
分类号:
TP274
DOI:
.
文献标志码:
A
摘要:
针对多传感器协同探测系统中系统偏差呈现出的随机性和突变性,以及估计方法的误差增加、估计鲁棒性较难保证等问题,在贝叶斯估计框架下,提出了基于凸优化的系统偏差估计方法。该方法首先依据最大似然估计准则推导出量测最大似然函数,并将其变形为与状态参数无关的多传感器量测最大似然函数;其次,结合系统偏差投影等式和待估参数范围不等式两类约束条件,将最大似然估计问题转化为具有目标函数、约束函数的凸优化问题;最后,利用拉格朗日乘子方法构造系统偏差二次函数,并在约束条件下利用凸优化技术实现多传感器系统偏差的优化求解。仿真结果表明,相比于同等条件下的其他方法,所提方法提高了估计精度,降低了时间复杂度。
Abstract:
Aiming at the stochastic and abruptness of systematic biases in multi-sensors collaborative detection system, and the increase of estimation error and the difficulty of ensuring robustnesss, a novel method of systematic biases estimation was presented based on convex optimization under the Bayesian framework. Firstly, the maximum likelihood function of measurement was derived under the optimum criterion of maximum likelihood. Then, the maximum likelihood function for multi-sensors measurements was given, and it was not concerned with state parameters. Secondly, combining the systematic biases projection equation and the parameter range inequality to be estimated, we translate the matter of maximum likelihood into the convex programming including objective function constraint function. Thirdly, the quadratic function of systematic biases was structured utilizing Lagrange multiplier method, subsequently, some optimization solutions of multi-sensors systematic biases were obtained using convex programming under constrain conditions. Simulation results showed that, comparing to other methods, the proposed algorithm effectively improved estimation precision, and reduced time complexity.

参考文献/References:

[1]Dana M P.Registration: a prerequisite for multiple sensor tracking[C]//Multitarget-multisensor Tracking: Advanced Applications. Norwood: Artech House, 1990: 155-185.
[2]熊伟,潘旭东,彭应宁,等.基于不敏变换的动基座传感器偏差估计方法[J].航空学报,2010,31(4): 819-824.
[3]李家强,赵春艳,陈金立,等.WGS-84坐标系下改进的广义最小二乘配准算法[J].上海交通大学学报,2016,50(5): 771-775.
[4]Zhang F H, Knoll A. Systematic error modeling and bias estimation [J].Sensors, 2016,16(5):729..
[5]Ye H, Yang Z P, Li J. BP neural networks based sensor registration algorithm[C]// Proceedings of 2011 International Conference on Computational and Information Sciences.Chengdu, China: IEEE,2011: 759-761.
[6]赵季红,谢志勇,曲桦,等.基于最大互相关熵UKF的传感网目标状态和系统偏差联合估计算法[J].计算机应用研究,2019,36(2):256-259.
[7]谢宇婷,周军,卢晓东.基于多导弹协同目标探测的空间配准算法[J].航空兵器,2018,304(2):11-17.
[8]iga Povalej. Quasi-Newton’s method for multiobjective optimization[J]. Journal of Computational and Applied Mathematics, 2014, 255:765-777.
[9]Dong X L, Liu H, He Y. A self-adjusting conjugate gradient method with sufficient descent condition and conjugacy condition [J]. Journal of Optimization Theory and Applications, 2015, 165(1):225-241.
[10]Wang Y, Ozbay B, Sharif M, et al.Convex optimization approaches to information structured decentralized control[J]. IEEE Transactions on Automatic Control,2018,63(10):3393-3403.
[11]邓先礼,魏波,曾辉,等.基于多种群的自适应迁移PSO算法[J].电子学报,2018,46(8):1858-1865.
[12]吕红芳,顾幸生. 基于蚁群神经网络的两级信息融合算法[J].上海交通大学学报,2016,50(8):1323-1330.
[13]王邑,孙金标,肖明清,等. 基于类型2区间模糊K近邻分类器的动态武器-目标分配方法研究[J]. 系统工程与电子技术,2016,38(6):1314-1319.
[14]司刚全,李水旺,石建全,等. 采用改进果蝇优化算法的最小二乘支持向量机参数优化方法[J].西安交通大学学报, 2017,51(6):14-19.
[15]周林,梁彦,潘泉.基于Metropolis-Hastings抽样的系统误差配准方法[J].系统工程与电子技,2012,34(3): 433-438.
[16]周林.多传感器系统偏差在线估计方法研究[D].西安:西北工业大学,2013.
[17]Stephen Boyd,Lieven Vandenberghe. Convex optimization[M]. UK: Cambridge University Press, 2004: 4-8.

备注/Memo

备注/Memo:
收稿日期:2019-01-25
基金项目:国家自然科学基金项目资助(61771006,61701170);河南省高等学校重点研究计划项目资助(19A413006);河南省科技发展计划项目资助(192102210254);河南大学一流学科培育项目资助(2018YLTD04)
作者简介:周林(1977—),女,河南开封人,博士,副教授,研究方向:多源信息融合,传感器管理。E-mail:zhoulin@henu.edu.cn.
更新日期/Last Update: 2019-09-12