Calibration of mems based inertial measurement unit using long short-Term memory network

Jinkui Wang*, Wenzhong Lou, Weitong Liu, Peng Liu

*Corresponding author for this work

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

4 Citations (Scopus)

Abstract

MEMS-based IMU performance is limited in some applications for its low accuracy, resulting from combine of many factors like orthogonal error, nonlinear scale factor, random drift, drift periodic error, etc. Calculating the main error matrix is one traditional solution, which can't express full error model of IMU. In this paper, MEMS based IMU's model has been developed, deep neural network was introduced to analyze the error model and the correction parameters were figured out. In the error model and calibration algorithm, output data of IMU are feed in Long Short-Term Memory Network(LSTM) with reference linear acceleration and angular rate, error character is delivered, and correction parameters are educed after training. Our experimental result on three-Axle turntable have proved the effectiveness of our algorithm.

Original languageEnglish
Title of host publicationProceedings of the 14th Annual IEEE International Conference on Nano/Micro Engineered and Molecular Systems, NEMS 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages118-121
Number of pages4
ISBN (Electronic)9781728116297
DOIs
Publication statusPublished - Apr 2019
Event14th Annual IEEE International Conference on Nano/Micro Engineered and Molecular Systems, NEMS 2019 - Bangkok, Thailand
Duration: 11 Apr 201914 Apr 2019

Publication series

NameProceedings of the 14th Annual IEEE International Conference on Nano/Micro Engineered and Molecular Systems, NEMS 2019

Conference

Conference14th Annual IEEE International Conference on Nano/Micro Engineered and Molecular Systems, NEMS 2019
Country/TerritoryThailand
CityBangkok
Period11/04/1914/04/19

Keywords

  • MEMS inertial measurement unit
  • error calibration
  • recurrent neural network

Fingerprint

Dive into the research topics of 'Calibration of mems based inertial measurement unit using long short-Term memory network'. Together they form a unique fingerprint.

Cite this