Maximum weight lifting prediction considering dynamic joint strength

Abstract

This thesis describes an efficient optimization method for predicting the maximum lifting weight considering dynamic joint strength in symmetric box lifting using a skeletal model. Dynamic joint strength is modeled as a three-dimensional function of joint angle and joint angular velocity based on experimentally obtained joint strength data. The function is further formulated as the joint torque limit constraint in an inverse dynamics optimization formulation to predict the lifting motion. In the proposed optimization formulation, external load is treated as design variables along with joint angle profiles, which are represented by control points of B-spline curves. By using this new formulation, dynamic lifting motion and strategy can be predicted for a symmetric maximum weight box lifting task with given initial and final box locations. Results show that incorporating dynamic strength is critical in predicting the lifting motion in extreme lifting conditions. The prediction outputs in joint space are incorporated in OpenSim software to find out muscles force and activity during the movement. Electromyography data are collected for a regular weight lifting to validate the integration process between the predictive model (joint model) and OpenSim model (muscle model). The proposed algorithm and analysis method based on motion prediction and OpenSim can be further developed as a useful ergonomic tool to protect workers from injury in manual material handling.