AbstractView references

In this research, we hypothesize that head-mounted accelerometric sensors can be effectively utilized in the tasks of motion activity classification and balance control evaluation, on a par with conventional methods based on stabilometric platforms. Following this hypothesis, we carried out a series of short experiments with stabilometric system "Stabilan-01-2" and consumer-grade tri-axial accelerometer. Collected motion signals were batch-processed and passed through feature extraction pipeline. After that, characteristic spaces were formed and classified using statistical machine learning methods. Results of the classification indicated that feature space of accelerometric data is informative enough to accurately classify motion activity associated with balance control, thus confirming our initial hypothesis. These results demonstrate that accelerometers can be used as a low-costly and portable alternative to stabilometric systems, and suggest a promising and novel approach to balance control assessment. © 2019 Author(s).