Fikri, Muhammad Naufal (2024) Membangun Model Regresi Untuk Identifikasi Kondisi Mengantuk Saat Berkendara Melalui Analisis Anomali Rentang Pola Gestur Tubuh Bagian Atas (Upper Limb). Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5010201080_Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5010201080_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (8MB) | Request a copy |
Abstract
Kecelakaan lalu lintas merupakan permasalahan global yang masih meningkat. Setidaknya 1,19 juta nyawa di dunia meregang di jalan raya akibat kecelakaan lalu lintas. Peningkatan 8% kecelakaan lalu lintas di Indonesia secara tahunan terus terjadi hingga akhir tahun 2022. Penyebab utama dari kecelakaan tersebut (93,5%) adalah kesalahan manusia. Tidak sedikit proporsi kelelahan dan kondisi mengantuk berada dalam salah satu penyebabnya. Sehingga pencegahan human error tersebut perlu dirumuskan. Penelitian ini hadir untuk mengetahui dan menganalisis kondisi mengantuk ketika berkendara menggunakan identifikasi pola perubahan gestur pada struktur segmen tulang bagian atas. Tiga macam sudut (leher, punggung, dan antara leher dengan punggung) ditinjau melalui dua penampang (frontal dan Sagital) yang didapatkan melalui motion capture PNS 3. Sistem identifikasi juga dihibridasi berdasarkan supporting validator (heart rate dan kedipan mata). Hubungan didapatkan melalui persamaan regresi logistik biner antara sudut postur dengan kondisi mengantuk (r2 = 90,02%). Devaisi maksimum didapatkan seperti sudut leher dengan sumbu vertikal (maksimum 14,84°± 16,17°), sudut punggung dengan sumbu vertikal (maksimum 7,51°± 7,42°), dan sudut antara leher dengan punggung (maksimum 76,75°±14,26°) untuk rerata sudut. Berdasarkan analisis keandalan, didapatkan nilai mean life sebesar 72 menit dan nilai warranty time sebesar 69,54 menit (ketika R(t) = 0,5).
=================================================================================================================================
Traffic accidents are a global problem that is still increasing. At least 1.19 million people worldwide lose their lives on the road due to traffic accidents. An 8% annual increase in traffic accidents in Indonesia is projected by the end of 2022. The main cause of these accidents (93.5%) is human error. A significant proportion of fatigue and drowsiness are among the causes. So the prevention of human error needs to be formulated. This research is presented to determine and analyze drowsy conditions when driving using the identification of gesture recognition patterns in the structure of the upper extremities’ segments. Three kinds of angles (neck, back, and between neck and back) are reviewed through two cross-planes (frontal and Sagittal) obtained through motion capture of PNS 3. The identification system was also hybridized using supporting validators (heart rate and eye blinking). The relationship was modeled through a binary logistic regression equation correlating posture angles with drowsiness conditions (r² = 90.02%). The maximum deviations were observed in the neck angle relative to the vertical axis (14.84° ± 16.17°), the back angle relative to the vertical axis (7.51° ± 7.42°), and the angle between the neck and back (76.75° ± 14.26°). Reliability analysis indicated a mean life of 72 minutes and a warranty time of 69.54 minutes (when R(t) = 0.5).
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Human Reliability, Human Behavior, Safety Driving, Drowsiness, Human Motion Analysis, Keandalan Manusia, Perilaku Manusia, Keselamatan Berkendara |
| Subjects: | H Social Sciences > HE Transportation and Communications > HE5614.2 Traffic safety |
| Divisions: | Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis |
| Depositing User: | Muhammad Naufal Fikri |
| Date Deposited: | 19 Jul 2024 08:39 |
| Last Modified: | 19 Jul 2024 08:39 |
| URI: | http://repository.its.ac.id/id/eprint/108528 |
Actions (login required)
 |
View Item |