Using Deep Learning for Motion Analysis of 3D Motion Capture Data for Forecasting Motion and Fatigue
| dc.contributor.advisor | Valles, Damian | |
| dc.contributor.author | Hernandez, Geovanni | |
| dc.contributor.committeeMember | Méndez, Francis A. | |
| dc.contributor.committeeMember | Jimenez, Jesus A. | |
| dc.date.accessioned | 2021-05-10T17:02:16Z | |
| dc.date.available | 2021-05-10T17:02:16Z | |
| dc.date.issued | 2021-05 | |
| dc.description.abstract | Industrial Revolution 4.0 is defined as the interconnection of Information, Communications Technologies (ICT) within the industry. In the occupation of laborers, stock, and material mover they are often subjected to repetitive motions that cause exhaustion (or fatigue) that could potentially lead to work-related musculoskeletal disorder (WMSD). The most common repetitive motions are lifting, pulling, pushing, carrying, and walking with load, which are also known as Manual Material Handling (MMH) operations. There has been work using a machine learning technique known as Recurrent Neural Network (RNN) to predict short and long-term motions from motion capture measurements but research in using motion capture data related to MMH to measure the fatigue needs exploration. For this research, only the lifting motion is considered. Motion data is collected as time-stamped motion data using infrared cameras at a rate of 100Hz of a subject performing repetitive lifting motion. The data is a combination of XYZ coordinates from 39 reflective markers. Along with motion data, the subject will self-report the perceived level of fatigue using the Borg scale every minute. All this data can be merged into one to further be used for analysis. Since motions occur over time for a duration of time, this data is used as input to a time-series deep learning technique known as Long Short-Term Memory and Gated Recurrent Unit models. Using these models, this research will evaluate the deep learning technique and motion capture data to perform motion analysis to forecast univariate motion data and to also predict the fatigue based on the displacement movement from each marker. | |
| dc.description.department | Engineering | |
| dc.format | Text | |
| dc.format.extent | 143 pages | |
| dc.format.medium | 1 file (.pdf) | |
| dc.identifier.citation | Hernandez, G. (2021). <i>Using deep learning for motion analysis of 3D motion capture data for forecasting motion and fatigue</i> (Unpublished thesis). Texas State University, San Marcos, Texas. | |
| dc.identifier.uri | https://hdl.handle.net/10877/13522 | |
| dc.language.iso | en | |
| dc.subject | Industry 4.0 | |
| dc.subject | Machine learning | |
| dc.subject | Deep learning | |
| dc.subject | Motion capture | |
| dc.subject | Manual material handling | |
| dc.subject.lcsh | Machine learning | |
| dc.subject.lcsh | Materials handling | |
| dc.title | Using Deep Learning for Motion Analysis of 3D Motion Capture Data for Forecasting Motion and Fatigue | |
| dc.type | Thesis | |
| thesis.degree.department | Engineering | |
| thesis.degree.discipline | Engineering | |
| thesis.degree.grantor | Texas State University | |
| thesis.degree.level | Masters | |
| thesis.degree.name | Master of Science |