Marcos Mazón, Daniel (2021) Convolutional Recurrent Neural Network: IMU-based Locomotion Intent and Gait Phase Prediction for Transfemoral Amputees. Master's Thesis / Essay, Artificial Intelligence.
|
Text
mAI_2021_MarcosMazonD.pdf Download (1MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming.pdf Restricted to Registered users only Download (100kB) |
Abstract
This paper focuses on the design of deep neural network architectures for the real-time prediction of locomotion modes, transitions and gait phases for ten healthy subjects and one osseointegrated transfemoral amputee by using inertial measurement units (IMU). Different neural network configurations are investigated by combining convolutional and recurrent layers. As input to the networks, the frequency aspect in the form of a spectrogram, of one IMU (located in the thigh) or two IMUs (located in both the thigh and the shank) are used. The system is able to predict seven different locomotion modes (sitting, standing, walking, ramp ascent and descent, stair ascent and descent), transitions among this locomotion modes and the gait phases corresponding to each of the locomotion modes. The results show that a system composed of CNN + LSTM networks is able to predict user intention with a mean F1-score of 0.893 and 0.910 for the healthy subjects, and 0.921 and 0.947 for the amputee subject, using one and two IMUs respectively with a 5-fold cross-validation.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Carloni, R. and Schomaker, L.R.B. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 27 Oct 2021 11:53 |
| Last Modified: | 17 Nov 2022 11:01 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/26227 |
Actions (login required)
 |
View Item |