Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Automated pre-impact fall detection
Thesis (MSc)--Stellenbosch University, 2024.
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Published: |
Stellenbosch : Stellenbosch University
2025
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613788793995264 |
|---|---|
| access_status_str | Open Access |
| author | Swanepoel, Mia |
| author2 | Coetzer, J. |
| author_browse | Coetzer, J. Swanepoel, Mia |
| author_facet | Coetzer, J. Swanepoel, Mia |
| author_sort | Swanepoel, Mia |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description |
Thesis (MSc)--Stellenbosch University, 2024. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/131920 |
| institution | Stellenbosch University (South Africa) |
| last_indexed | 2026-06-10T12:41:42.984Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/131920 Automated pre-impact fall detection Swanepoel, Mia Coetzer, J. Stellenbosch University. Faculty of Science. Dept. of Applied Mathematics. Support vector machines Wearable technology Accidental falls -- Prevention Convolutions (Mathematics) UCTD Thesis (MSc)--Stellenbosch University, 2024. ENGLISH ABSTRACT: The global population is aging rapidly, with individuals aged 60 and older now outnum- bering children under five. As a result, age-related injuries, including falls, have become increasingly common. Falls are the second leading cause of injury and death in the elderly. Systems such as cameras and wearable devices have been developed to monitor movement and detect falls. Research is now focusing on detecting falls before impact to enable possible interventions, such as wearable airbags, to reduce injuries. Current systems use methods like thresholding techniques and machine learning models to detect falls before they occur. This study examines various models that utilise time series data from a wearable device to detect falls before impact. The models include a thresholding technique, a support vector machine (SVM), a convolutional neural network (CNN), a convolutional long short term memory (ConvLSTM) network, a transformer, and the state of the art iTransformer. The data used in this study to train and test the models is the KFall dataset which includes 2729 activities of daily living (ADLs) and 2346 falls. The ConvLSTM model achieved the longest lead time of 391 ± 109 ms and the largest area under the receiver operating characteristic curve (AUC) of 0.94 as well as the best sensitivity and specificity trade off. The iTransformer showed promising results with an AUC of 0.84 and lead time of 375 ± 111 ms as well as a notable specificity of 91.22%. This research demonstrates the capability of a thresholding technique and deep learning models to accurately detect falls before impact in order to improve reaction time and reduce fall-related injuries. AFRIKAANSE OPSOMMING: Die gemiddelde ouderdom van mense wêreldwyd neem jaarliks toe. Daar is tans meer individue van 60 jaar en ouer as wat daar kinders onder die ouderdom van 5 is. Dit lei tot ′n verhoging van beserings wat verband hou met ouderdom, soos bejaardes wat val, wat die tweede grootste oorsaak van beserings en sterftes onder bejaardes is. Stelsels soos kameras en draagbare toestelle is ontwikkel om beweging te monitor en valle te bespeur. Navorsing fokus tans op bespeuring van valle voor impak om moontlike ingrypings, soos draagbare lugsakke, te maak om beserings te verminder. Huidige stelsels gebruik metodes soos drempelwaardes en masjienleer-modelle om valle voor impak te bespeur. Hierdie studie ondersoek verskeie modelle wat tydreeksdata van ’n draagbare toestel gebruik om valle voor impak te bespeur. Die modelle sluit ’n drempeltegniek, ’n ondersteuningsvektor masjien (SVM), ’n konvolusionele neurale netwerk (CNN), ’n konvolusionele lang-korttermyngeheunetwerk (ConvLSTM), ’n transformator en die gesofistikeerde iTransformeerder in. Die data wat in hierdie studie gebruik is om die modelle af te rig en te toets, is die KFall-dataset, wat 2729 aktiwiteite van alle daagse lewe (ADL’s) en 2346 valle insluit. Die ConvLSTM-model het die langste lei-tyd van 391 ± 109 ms en die grootste area onder die ontvanger werkskenmerk kromme (AUC) van 0.94 behaal, sowel as die beste balans tussen sensitiwiteit en spesifisiteit. Die iTransformeerder het belowende resultate getoon met ’n AUC van 0.84 en lei-tyd van 375 ± 111 ms, sowel as ’n noemenswaardige spesifisiteit van 91.22%. Hierdie navorsing demonstreer die vermoë van ’n drempeltegniek en diep leer-modelle om valle akkuraat voor impak te bespeur ten einde reaksietyd te verbeter en valverwante beserings te verminder. Masters 2025-04-09T14:16:13Z 2025-04-09T14:16:13Z 2024-12 Thesis https://scholar.sun.ac.za/handle/10019.1/131920 Stellenbosch University xii, 61 pages : illustrations application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Support vector machines Wearable technology Accidental falls -- Prevention Convolutions (Mathematics) UCTD Swanepoel, Mia Automated pre-impact fall detection |
| title | Automated pre-impact fall detection |
| title_full | Automated pre-impact fall detection |
| title_fullStr | Automated pre-impact fall detection |
| title_full_unstemmed | Automated pre-impact fall detection |
| title_short | Automated pre-impact fall detection |
| title_sort | automated pre impact fall detection |
| topic | Support vector machines Wearable technology Accidental falls -- Prevention Convolutions (Mathematics) UCTD |
| url | https://scholar.sun.ac.za/handle/10019.1/131920 |
| work_keys_str_mv | AT swanepoelmia automatedpreimpactfalldetection |