RNN-aided Human Velocity Estimation from a Single IMU

Feigl T, Kram S, Woller P, Siddiqui RH, Philippsen M, Mutschler C (2020)

Publication Language: English

Publication Type: Journal article, Original article

Publication year: 2020

Journal

Sensors MDPI

Book Volume: 13

Pages Range: 1-31

Article Number: 3656

Journal Issue: 20

URI: https://www.mdpi.com/1424-8220/20/13/3656

DOI: 10.3390/s20133656

Open Access Link: https://res.mdpi.com/d_attachment/sensors/sensors-20-03656/article_deploy/sensors-20-03656.pdf

Abstract

Pedestrian Dead Reckoning (PDR) uses inertial measurement units (IMUs) and combines velocity and orientation estimates to determine a position. The estimation of the velocity is still challenging, as the integration of noisy acceleration and angular speed signals over a long period of time causes large drifts. Classic approaches to estimate the velocity optimize for specific applications, sensor positions, and types of movement and require extensive parameter tuning. Our novel hybrid filter combines a convolutional neural network (CNN) and a bidirectional recurrent neural network (BLSTM) (that extract spatial features from the sensor signals and track their temporal relationships) with a linear Kalman filter (LKF) that improves the velocity estimates. Our experiments show the robustness against different movement states and changes in orientation, even in highly dynamic situations. We compare the new architecture with conventional, machine, and deep learning methods and show that from a single non-calibrated IMU, our novel architecture outperforms the state-of-the-art in terms of velocity (≤0.16m/s) and traveled distance (≤3m/km). It also generalizes well to different and varying movement speeds and provides accurate and precise velocity estimates.

Authors with CRIS profile

Tobias Feigl Lehrstuhl für Informatik 2 (Programmiersysteme) Sebastian Kram Lehrstuhl für Informationstechnik mit dem Schwerpunkt Kommunikationselektronik (Stiftungslehrstuhl) (LIKE) Michael Philippsen Lehrstuhl für Informatik 2 (Programmiersysteme) Christopher Mutschler

Related research project(s)

Recurrent Neuronal Networks (RNNs) for Real-Time Estimation of Nonlinear Motion Models (RuNN) Oct. 1, 2017 - March 31, 2021

Involved external institutions

Fraunhofer-Institut für Integrierte Schaltungen (IIS) DE Germany (DE)

How to cite

APA:

Feigl, T., Kram, S., Woller, P., Siddiqui, R.H., Philippsen, M., & Mutschler, C. (2020). RNN-aided Human Velocity Estimation from a Single IMU. Sensors, 13(20), 1-31. https://doi.org/10.3390/s20133656

MLA:

Feigl, Tobias, et al. "RNN-aided Human Velocity Estimation from a Single IMU." Sensors 13.20 (2020): 1-31.

BibTeX: Download