Improvement of a Forward Dynamic MPC Based Human Gait Model

This paper develops an improvement to an existing forward dynamic human gait model. A human gait model was developed previously to assist virtual testing prostheses and orthoses. The model consists of a plant model and a controller model. The central tenet to the model is the model predictive control (MPC) algorithm, which is a highly robust controller. In the previous model, however, there are several drawbacks. First, the anthropometric and mechanical parameters in the parts of the model are specific to one person. Second, the simulation result of ground reaction force (GRF) is not realistic. In this paper, the anthropometric parameters are calculated based on commonly used models that approximate an average person’s size. As for the mechanical parameters, the spring and damper coefficients in the human joints and ground reaction force (GRF) system are estimated by using the parameter estimation module in MATLAB based on the experimental subject data. The paper concludes with a simulation results between the new improved model and the previous developed model.