Motion prediction and validation considering perceived risk-based three-dimensional collision avoidance | JoVE Visualize

Area of Science:

Biomechanics and Human Motion Analysis
Cognitive Science and Risk Perception
Robotics and Human-Computer Interaction

Background:

Predicting human upper extremity motion is crucial for digital human modeling, especially in tasks involving obstacle avoidance.
Existing models often lack the integration of cognitive factors like perceived risk, limiting their predictive accuracy.
Understanding how humans adjust movements based on perceived risk is key to improving safety and efficiency in human-centric applications.

Purpose of the Study:

To develop and validate an optimization-based motion prediction framework for human upper extremity movement in 3D collision avoidance tasks.
To integrate biomechanical constraints with cognitive perceived risk using Bayesian Decision Theory into the motion prediction model.
To advance digital human modeling by incorporating perceived risk beyond traditional artificial contact spheres.

Main Methods:

Collected experimental data of subjects performing reaching tasks with 3D obstacles of varying properties.
Developed an optimization formulation using B-Spline curves for joint angles, minimizing joint displacement and maximizing end-effector velocity.
Incorporated perceived risk-related constraints into the optimization framework, modifying a baseline model that excluded cognitive factors.

Main Results:

The inclusion of perceived risk significantly improved the prediction of minimum clearance distances around obstacles.
Human subjects exhibited greater clearance distances around fragile objects, demonstrating cautious behavior influenced by perceived risk.
The prediction method was validated by comparing simulated joint angle profiles with experimental data, showing good agreement.

Conclusions:

Integrating perceived risk into motion prediction frameworks significantly enhances the accuracy of digital human models for collision avoidance tasks.
This cognitive-informed approach offers a more realistic representation of human motion, moving beyond purely biomechanical or geometric constraints.
The findings have broad applications in ergonomics, rehabilitation, and human-robot interaction, improving workspace design, safety protocols, and collaborative robotics.