Optimization-based multibody dynamics modeling method for team lifting simulation

Abstract

A novel optimization-based multibody dynamics modeling method is proposed for two-dimensional (2D) team lifting prediction. The box itself is modeled as a floating-base rigid body in Denavit-Hartenberg representation. The interactions between humans and box are modeled as a set of grasping forces which are treated as unknowns (design variables) in the optimization formulation. An inverse dynamics optimization is used to simulate the team lifting motion where the dynamic effort of two humans is minimized subjected to physical and task-based constraints. The design variables are control points of cubic B-splines of joint angle profiles of two humans and the box, and the grasping forces between humans and the box. Analytical sensitivities are derived for all constraints and objective functions, including the varying unknown grasping forces. Two numerical examples are successfully simulated: one is to lift a 10 Kg box with the center of mass (COM) in the middle, and the other is the same weight box with the COM off the center. The humans' joint angle, torque, ground reaction force, and grasping force profiles are reported. The optimal solution is obtained in 151.99 seconds. The simulated motions are validated against the experimental joint angle profiles. Reasonable team lifting motion, kinematics, and kinetics are predicted using the proposed novel multibody dynamic modeling approach and optimization formulation.