Adaptive Attitude Estimation Using a Hybrid Model-Learning Approach

Eran Vertzberger, Itzik Klein

Research output: Contribution to journalArticlepeer-review


Attitude determination using the smartphone's inertial sensors poses a major challenge due to the sensor's low-performance grade and the varying nature of the walking pedestrian. In this article, data-driven techniques are employed to address that challenge. To that end, a hybrid deep learning (DL) and model-based solution for attitude estimation is proposed. Here, classical model-based equations are applied to form an adaptive complementary filter (CF) structure. Instead of using constant or model-based adaptive weights, the accelerometer weights in each axis are determined by a unique neural network (NN). The performance of the proposed hybrid approach is evaluated relative to popular model-based approaches using experimental data.

Original languageEnglish
Article number8505009
JournalIEEE Transactions on Instrumentation and Measurement
StatePublished - 2022

Bibliographical note

Publisher Copyright:
© 1963-2012 IEEE.


  • Attitude and heading reference system (AHRS)
  • complementary filter (CF)
  • data-driven navigation
  • deep learning (DL)
  • inertial sensors

ASJC Scopus subject areas

  • Instrumentation
  • Electrical and Electronic Engineering


Dive into the research topics of 'Adaptive Attitude Estimation Using a Hybrid Model-Learning Approach'. Together they form a unique fingerprint.

Cite this