Towards a Multifidelity Digital Human Model for Use in Simulation Environments for Tactile Human/Robot Interactions.

Area of Science:

• Robotics and Autonomous Systems (RAS)
• Human-Robot Interaction (HRI)
• Computational Biomechanics

Background:

• Increasing interest in RAS for battlefield casualty extraction and medical interventions.
• Current digital human models (DHMs) lack biofidelity and scalability for robotic simulations.
• Need for accurate human modeling in robotic manipulation to prevent further injury during Human-Robot Interaction (HRI).

Purpose of the Study:

• To develop a multifidelity digital human model (M-DHM) suitable for robotic simulation environments.
• To integrate low-fidelity multibody dynamics (MD) DHMs with high-fidelity finite-element (FE) DHMs.
• To provide path planners with both point and distributed loading data during HRI events.

Main Methods:

• Developed a software architecture integrating two low-fidelity MD-DHMs and one high-fidelity FE-DHM.
• Enabled exchange of human joint kinematics and applied forces between MD-DHM(s) and FE-DHM.
• Developed methods to retrieve multifidelity anatomical loading (M-AL) and estimate it using reduced-order models.

Main Results:

• Successfully demonstrated the M-DHM for grasping and palpation of lower arms.
• Validated the model for repositioning and dragging of a leg in simulation.
• Showcased the model's capability to provide comprehensive loading data for HRI planning.

Conclusions:

• The M-DHM advances next-generation planning algorithms for robotic applications.
• Potential for military applications, such as search-and-rescue missions.
• Future civilian applications include RAS for patient manipulation and healthcare.