计算机技术与发展2018，Vol.28Issue(5)：5-8,4.DOI:10.3969/j.issn.1673-629X.2018.05.002
基于最小二乘法和BP神经网络的TOA定位算法
A TOA Positioning Algorithm Based on Least Square Method and BP Neural Network
摘要
Abstract
In view of the problem of localization with low accuracy caused by complex indoor environment,we propose a new localization algorithm based on time of arrival (TOA).The linear error term is introduced in the least square method,and the BP neural network training is used to create the ranging model instead of the traditional scheme,so as to eliminate the over-experience dependence on the en-vironment,improve the universality of the algorithm for different environments.In this paper,we select the root mean square error as the performance evaluation criterion,by which we perform a comparative testing of the proposed method and the two representative algo-rithms,traditional least squares method and classical Chan algorithm.By testing more than 6000 records in five different environments,the results show that the improved algorithm is able to obtain the better performance than the Chan algorithm and the least squares estimation method,and enhances the localization accuracy greatly.In addition,cross-experiments show that the models in different scenarios is uni-versal.关键词
到达时间/最小二乘法/BP神经网络/交叉实验Key words
TOA/least square estimation/BP neural network/cross experiment分类
信息技术与安全科学引用本文复制引用
浦佳祺,陈德旺..基于最小二乘法和BP神经网络的TOA定位算法[J].计算机技术与发展,2018,28(5):5-8,4.基金项目
福建省闽江学者科研启动项目(510146) （510146）
工信部倍增计划(00101522) （00101522）
福州大学人才引进科研启动项目(0030-510206) （0030-510206）
