Javascript must be enabled for the correct page display

Deep reinforcement learning for physics-based musculoskeletal simulations of healthy subjects and transfemoral prosthesis users during normal walking

Vree, Leanne de (2020) Deep reinforcement learning for physics-based musculoskeletal simulations of healthy subjects and transfemoral prosthesis users during normal walking. Bachelor's Thesis, Artificial Intelligence.

[img] Text
Bachelor_Thesis_s3002195_Leanne.pdf

Download (1MB)
[img] Text
toestemming.pdf
Restricted to Registered users only

Download (105kB)

Abstract

: This paper focuses on the implementation of a Deep Reinforcement Learning algorithm for the simulation of physics-based musculoskeletal models of both healthy subjects and transfemoral prostheses’ users during normal level ground walking. The algorithm is based on the Proximal Policy Optimization approach in combination with reward shaping and imitation learning to reduce the computation time of the training while guaranteeing a natural walking gait. Firstly, the optimization algorithm is designed for the OpenSim model of a healthy subject and validated with experimental data. Afterwards, the optimization algorithm is applied to the OpenSim model of a transfemoral prosthesis’ user which has been obtained by substituting the healthy muscles with muscle-like linear motors. The model of the transfemoral prosthesis’ user shows a stable gait, with a forward dynamic comparable to the healthy subject model, yet using higher muscle forces.

Item Type: Thesis (Bachelor's Thesis)
Supervisor name: Carloni, R.
Degree programme: Artificial Intelligence
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 16 Sep 2020 09:44
Last Modified: 11 Jun 2021 12:35
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/23233

Actions (login required)

View Item View Item