A Method for Constructing a Dynamic Seat Adjustment Model to Mitigate Prolonged-Sitting Discomfort

Area of Science:

• Human-Computer Interaction
• Ergonomics
• Automotive Engineering

Background:

• Prolonged sitting in public transport causes discomfort.
• Existing dynamic seating solutions lack personalization and coordinated adjustments.
• Behavioral and anthropometric variability are often overlooked in seat design.

Purpose of the Study:

• To develop a dynamic seat-adjustment model considering user behavior and height.
• To generate coordinated, low-frequency seat trajectories for reduced discomfort.
• To preliminarily validate the model's effectiveness in public transport settings.

Main Methods:

• A multi-output random forest model was trained on participant-adjusted seat configurations.
• Data included elapsed sitting time, user behavior, and height-based strategy groups.
• A within-subject crossover experiment compared Static, Mean-trajectory, and Dynamic seating conditions.

Main Results:

• Preferred seat configurations varied by behavior and height group, changing over time.
• The dynamic model generated coordinated trajectories, outperforming independent adjustments.
• Dynamic seating significantly reduced overall discomfort compared to static seating.
• Additional comfort benefits were observed in neck and trunk regions with dynamic seating.

Conclusions:

• A stratified multi-output model effectively creates coordinated, time-varying seat trajectories.
• This approach lowers subjective discomfort during prolonged sitting.
• The framework provides a basis for passive, low-frequency adjustments in transportation seats.