Deep Reinforcement Learning for Physics-Based Musculoskeletal Simulations of Transfemoral Prosthesis' Users during the Transition between Normal Walking and Stairs Ascending

Petrescu, Ruxandra (2021) Deep Reinforcement Learning for Physics-Based Musculoskeletal Simulations of Transfemoral Prosthesis' Users during the Transition between Normal Walking and Stairs Ascending. Bachelor's Thesis, Artificial Intelligence.

Preview	Text Bachelor's Thesis Ruxandra Petrescu s3610187.pdf Download (1MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (104kB)

Abstract

This paper proposes to use deep reinforcement learning for the simulation of a physicsbased musculoskeletal model of transfemoral prostheses’ users during the transition between levelground walking and stairs ascend. The deep reinforcement learning algorithm uses the proximal policy optimization with covariance matrix adaptation and imitation learning to guarantee the level ground walking and the ascension of stairs. The optimization algorithm is implemented for the OpenSim model of transfemoral prosthesis’ users. The transfemoral prosthesis has two actuators for the knee joint and the ankle joint. This study shows that the model can only take a step while normal walking, starting with the prosthesis, and it has not been able to take the second step, the transition to ascending the stairs, even with the increase of 50% of the maximum muscles’ force.

Item Type:	Thesis (Bachelor's Thesis)
Supervisor name:	Carloni, R. and Raveendranathan, V.
Degree programme:	Artificial Intelligence
Thesis type:	Bachelor's Thesis
Language:	English
Date Deposited:	29 Jul 2021 08:42
Last Modified:	10 Jan 2024 13:12
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/25500

Actions (login required)

View Item