Sistem Pemantauan Postur Duduk Pada Anak Cerebral Palsy Menggunakan Multi-Sensor Berbasis Machine Learning Pada Kursi Roda

Izzah, Calia Jahidatul (2026) Sistem Pemantauan Postur Duduk Pada Anak Cerebral Palsy Menggunakan Multi-Sensor Berbasis Machine Learning Pada Kursi Roda. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023221034-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Cerebral palsy (CP) merupakan gangguan perkembangan motorik yang menyebabkan anak kesulitan mengendalikan gerakan dan postur tubuh. Anak dengan Gross Motor Function Classification System (GMFCS) level III–V umumnya sangat bergantung pada kursi roda untuk mobilitas sehari-hari, dan penggunaan kursi roda jangka panjang berisiko menimbulkan asimetri postural, nyeri kronis, dan deformitas tulang belakang. Penelitian terdahulu telah mengembangkan sistem pemantauan postur dengan sensor tunggal seperti Force Sensitive Resistor (FSR), Inertial Measurement Unit (IMU), atau kamera, namun masing-masing masih memiliki keterbatasan akurasi, ketergantungan lingkungan, dan informasi yang terbatas. Penelitian ini mengusulkan sistem pemantauan postur duduk anak CP pada kursi roda berbasis multi-sensor yang mengintegrasikan IMU, FSR, dan kamera, dengan data ketiga sensor diproses menggunakan machine learning untuk mengklasifikasikan lima postur duduk: tegak, condong kanan, condong kiri, condong depan, dan condong belakang. Hasil pengujian menunjukkan sistem mencapai akurasi testing rata-rata 91,61% menggunakan kombinasi 22 fitur, lebih tinggi dibandingkan penggunaan seluruh 26 fitur dengan akurasi sebesar 87,58%, sekaligus dengan waktu eksekusi tercepat. Sistem ini mampu memberikan informasi postur yang lebih komprehensif dibandingkan sensor tunggal serta mendukung deteksi dini asimetri postur. Dengan demikian, sistem ini berpotensi dikembangkan sebagai teknologi pemantauan berbasis sensor untuk pencegahan deformitas dan peningkatan kualitas hidup anak penyandang cerebral palsy.
=====================================================================================================================================
Cerebral palsy (CP) is a motor developmental disorder that causes difficulty in controlling body movement and posture. Children with Gross Motor Function Classification System (GMFCS) level III–V are generally highly dependent on wheelchairs for daily mobility, and prolonged wheelchair use carries the risk of postural asymmetry, chronic pain, and spinal deformity. Previous studies have developed posture monitoring systems using single sensors such as Force Sensitive Resistors (FSR), Inertial Measurement Units (IMU), or cameras, yet each still has limitations in accuracy, environmental dependency, and information scope. This study proposes a multi-sensor sitting posture monitoring system for wheelchair-bound children with CP, integrating IMU, FSR, and camera sensors. Data from the three sensors are processed using machine learning to classify five sitting postures: upright, leaning right, leaning left, leaning forward, and leaning backward. Testing results show that the system achieves an average testing accuracy of 91.61% using a combination of 22 features, only using FSR and camera angles, excluding IMU head-angle features, higher than when using all 26 features with an average accuracy 87.58%, while also achieving the fastest execution time. The system provides more comprehensive postural information than single-sensor approaches and supports early detection of postural asymmetry. Accordingly, this system has the potential to be developed as a sensor-based monitoring technology for deformity prevention and improved quality of life for children with cerebral palsy.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Cerebral Palsy, FSR, IMU, Kamera, Kursi Roda, Machine Learning, Pemantauan Postur Camera, Cerebral Palsy, FSR, IMU, Machine Learning, Posture Monitoring, Wheelchair
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Calia Jahidatul Izzah
Date Deposited:	03 Aug 2026 01:45
Last Modified:	03 Aug 2026 01:45
URI:	http://repository.its.ac.id/id/eprint/141201

Actions (login required)

View Item