AI 기반 에어웹 구조 시트 쿠션의 골반 체압 분산 및 바른 자세 유지를 위한 최적화 연구

AI-Based Optimization of Pelvic Pressure Distribution and Upright Postural Support in Air-web Structure Seat Cushions

초록

Background: Prolonged sitting concentrates a mechanical load on the ischial region and may affect pelvic alignment, local discomfort, and musculoskeletal burden. However, seat-cushion evaluation has often relied on single-pressure metrics such as mean pressure or contact area, which are insufficient to explain the overall balance of pelvic support. Objectives: This study aimed to evaluate pelvic support in air-web material seat cushions using an objective pressure-based index and then to estimate the optimal firmness zone through regression and machine- learning analysis. Methods: Thirty healthy adults completed a repeated-measures experiment. Seated body-pressure was recorded with a 32×32 Tactilus sensor under five density and two backrest conditions. We defined a Pelvic Support Index (PSI) integrating four features: smoothness, thigh-to-ischial load redistribution, ischial peak- pressure control, and contact engagement. Density effects were assessed using quadratic regression and high- support classification. Subjective preference and backrest effects were analyzed separately. Results: In the seat-only condition, Cushion 3 with a density of 70.1 showed the highest mean PSI, 933.8±229.5. Quadratic regression revealed an inverted-U relationship between density and PSI and estimated the theoretical optimum at 79.6, with R =0.9667. Five machine learning models showed a strong performance for the high-support classification, with ROC-AUC values ranging from 0.9650 to 0.9879. Subjectively, Cushion 3 was selected as the first-choice cushion by eleven of 30 participants, 36.7%. However, individual- level agreement between PSI-based ranking and subjective ranking was limited, with Spearman p=–0.017. Conclusions: The proposed PSI provides an interpretable and objective framework for evaluating pelvic support from seated pressure distribution. An intermediate density zone centered around 70.1 to 79.6 showed the most favorable balance between ischial pressure control, anterior load redistribution, and overall pressure. Objective pelvic support did not fully match short-term subjective preference, and seat-with-backrest results should be interpreted separately from seat-only results.

키워드

Artificial intelligencebody pressure distributionpelvic supportseat cushion
제목
AI 기반 에어웹 구조 시트 쿠션의 골반 체압 분산 및 바른 자세 유지를 위한 최적화 연구
제목 (타언어)
AI-Based Optimization of Pelvic Pressure Distribution and Upright Postural Support in Air-web Structure Seat Cushions
저자
곽대훈함승헌주재걸
DOI
10.5668/JEHS.2026.52.3.270
발행일
2026-06
유형
Y
저널명
한국환경보건학회지
52
3
페이지
270 ~ 280