Posture prediction and optimization for a manual assembly operation involving lifting of weights

Abstract

The present work combines ergonomics with the posture prediction in the assembly process to avoid musculoskeletal issues of human operator. For improved productivity the operator should be in a better work environment and in sound health. The purpose of this paper is to provide a different perspective to avoid ergonomic risk factors in manual assembly. Here, a human is modeled as 20-DOF as modeled in robotic analysis and simulated in a virtual environment. In the present study, two objective cost functions i.e. joint discomfort function and energy expenditure function have been employed for evaluating the optimized posture. For posture prediction, a combined multi-objective optimization (MOO) method is used and the objective cost functions are minimized i.e. less joint discomfort and less energy in MOO method required to do the manual assembly operation and consequently, the results are compared and finally the movements are tested using REBA technique.

A Survey of Human Action Recognition and Posture Prediction

“…Liu et al [192] focused on streaming 3D skeleton sequences, and proposed dilated convolutional network for online action prediction. Rout et al [193] used posture analysis and mathematical modeling of the position of adjacent joints of muscles, so that this method could predict and optimize the posture of weight lifting assembly operations. Therefore, we introduce skeleton-based human posture prediction methods at first.…”