OPTIMIZATION-BASED THREE-DIMENSIONAL SYMMETRIC TOSSING MOTION PREDICTION AND COMPARISON WITH TWO-DIMENSIONAL MODEL

Weight tossing is a combined manual material handling task performed in human workplaces. The combined motion reduces labor time but comes at the cost of increased physical strain. The accumulated muscle fatigue may cause injuries, leading to high healthcare expenses. After the pandemic, workplace injuries have been in the spotlight with the trend of rising e-commerce. It is critical to analyze tossing tasks to suggest a solution to prevent injury risks. Tossing motion prediction provides insights into the human biodynamic system without real experimental data compared to traditional experimental-based methods. In addition, motion prediction can provide predicted joint torques needed for the tossing motion, and this information is useful to prevent potential injuries. This paper introduces an optimization-based 3D symmetric motion prediction and compares it with the 2D results. The optimization formulation of the 3D prediction is introduced. Lab experimental data of tossing a sandbag to a target point are collected from a healthy male subject and used to validate the predicted results. Results of the predicted joint angles, torques, and ground reaction forces (GRFs) from the 3D model are compared with those from the 2D model. The pros and cons of the 3D and 2D prediction models are reported, considering the higher computational cost of 3D prediction than the 2D. Finally, the future research directions of tossing motion are summarized.