An optimized fusion positioning algorithm based on BP neural network

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Multi-source fusion positioning is an effective approach when the positioning accuracy is unsatisfactory. To achieve accurate and continuous pedestrian positioning in complex indoor environments, we propose an optimized fusion positioning algorithm based on BP neural network. Firstly, the average positioning error of Wi-Fi is utilized to constrain the geomagnetic matching range. Secondly, Particle Swarm Optimization (PSO) is used to optimize the BP-AdaBoost ensemble learning algorithm, then the optimized BP-AdaBoost-PSO is used to fuse Wi-Fi positioning results and constrained geomagnetic positioning results. Simulation results indicate that the constraint method can reduce the positioning error caused by the geomagnetic mismatch, and the proposed fusion positioning algorithm can reduce the average execution time and improve the positioning accuracy.

Original languageEnglish
Title of host publicationProceedings of 2021 IEEE International Conference on Emergency Science and Information Technology, ICESIT 2021
EditorsGuorong Chen
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages244-248
Number of pages5
ISBN (Electronic)9781665435314
DOIs
StatePublished - 2021
Event2021 IEEE International Conference on Emergency Science and Information Technology, ICESIT 2021 - Chongqing, China
Duration: 22 Nov 202124 Nov 2021

Publication series

NameProceedings of 2021 IEEE International Conference on Emergency Science and Information Technology, ICESIT 2021

Conference

Conference2021 IEEE International Conference on Emergency Science and Information Technology, ICESIT 2021
Country/TerritoryChina
CityChongqing
Period22/11/2124/11/21

Keywords

  • ensemble learning
  • indoor positioning
  • multi-source fusion
  • neural network
  • particle swarm optimization

Fingerprint

Dive into the research topics of 'An optimized fusion positioning algorithm based on BP neural network'. Together they form a unique fingerprint.

Cite this