Learning and generating human natural behaviours for design evaluation using artificial neural networks

Area of Science:

• Biomechanics
• Human-Computer Interaction
• Machine Learning

Background:

• Current product design tools lack natural human motion animation.
• Existing strength prediction methods are often static and population-specific.
• There's a need for more flexible and dynamic biomechanical analysis tools.

Purpose of the Study:

• To develop a novel approach using neural networks for natural human behavior generation.
• To propose neural networks for adaptable biomechanical strength prediction.
• To enhance the flexibility of biomechanical analysis in product design.

Main Methods:

• Utilized neural networks for learning and generating human movement sequences.
• Applied neural networks to strength prediction, allowing for varied inputs and outputs.
• Focused on creating naturalistic animation and non-linear relationship mapping for strength.

Main Results:

• Demonstrated the capability of neural networks to learn and generate natural human behaviors.
• Showcased the flexibility of neural networks in strength prediction models.
• Enabled inclusion of dynamic strength and diverse population data.

Conclusions:

• Neural networks offer a powerful solution for creating natural human animations in design.
• The proposed method significantly improves the adaptability and accuracy of biomechanical strength prediction.
• This approach advances biomechanical analysis for more realistic product design and evaluation.