Estimation of Full-Body Poses Using Only Five Inertial Sensors: An Eager or Lazy Learning Approach?

Frank Wouda,Giovanni Bellusci,Peter Veltink,Matteo Giuberti

doi:10.3390/s16122138

Frank Wouda, Giovanni Bellusci + Show 2 more

Open Access

https://doi.org/10.3390/s16122138

Copy DOI

Abstract

Human movement analysis has become easier with the wide availability of motion capture systems. Inertial sensing has made it possible to capture human motion without external infrastructure, therefore allowing measurements in any environment. As high-quality motion capture data is available in large quantities, this creates possibilities to further simplify hardware setups, by use of data-driven methods to decrease the number of body-worn sensors. In this work, we contribute to this field by analyzing the capabilities of using either artificial neural networks (eager learning) or nearest neighbor search (lazy learning) for such a problem. Sparse orientation features, resulting from sensor fusion of only five inertial measurement units with magnetometers, are mapped to full-body poses. Both eager and lazy learning algorithms are shown to be capable of constructing this mapping. The full-body output poses are visually plausible with an average joint position error of approximately 7 cm, and average joint angle error of 7. Additionally, the effects of magnetic disturbances typical in orientation tracking on the estimation of full-body poses was also investigated, where nearest neighbor search showed better performance for such disturbances.

Highlights

Human motion has been a research topic of interest in many fields for a long time
The introduction of inertial motion capture systems, which do not rely on any external infrastructure, made full-body movement analysis feasible in an ambulatory setting [1]
We evaluate the impact of sensor placement (Section 3.1) and different activities (Section 3.2) on the performance of Nearest Neighbor Search (NNS) and Artificial Neural Networks (ANNs)

Summary

Introduction

Human motion has been a research topic of interest in many fields for a long time. The increasing availability of high-quality motion capture systems [1,2,3,4] contributed to this topic, by allowing easier and more accurate three-dimensional human motion capturing [5]. The introduction of inertial motion capture systems, which do not rely on any external infrastructure, made full-body movement analysis feasible in an ambulatory setting [1]. These systems require sensors to be attached to each main body segment (e.g., 17 sensors in Xsens MVN [6]). By reducing the number of required body-worn sensors, such systems would be less obtrusive and the usability would improve, which could potentially lead to applications that require use in daily life. Another probable benefit would be the reduction in costs

Methods

Results

Conclusion