Predict human body indentation lying on a spring mattress using a neural network approach

Area of Science:

• Biomechanics
• Ergonomics
• Materials Science

Background:

• Designing effective spring mattresses requires understanding human spinal interaction.
• Current mattress design often lacks personalized support based on individual spinal geometry.
• Accurate prediction of spinal indentation is crucial for ergonomic mattress development.

Purpose of the Study:

• To develop a predictive model for human body-mattress interaction.
• To optimize the design of five-zone spring mattresses for effective spinal support.
• To propose a practical design process for personalized mattress stiffness.

Main Methods:

• Development of a three-layer artificial neural network model.
• Simulation and prediction of human spinal indentation curves.
• Analysis of key spinal parameters: lumbar lordosis depth and inclination angles (cervicothoracic, thoracolumbar, lumbosacral, back-hip angle β).

Main Results:

• The artificial neural network accurately predicted spinal indentation curves.
• Optimal evaluation parameters were identified for mattress design.
• A method for designing five-zone spring mattresses with effective body support was established.
• Specific stiffness proportions were proposed for Chinese young women's body types.

Conclusions:

• The developed method is feasible and practical for predicting human-mattress interaction.
• The approach facilitates the design of ergonomically superior spring mattresses.
• This research contributes to personalized comfort and support in bedding design.