Synthesizing Get-Up Motions for Physics-based Characters

Description:

This project consists to learn a physics based character to get-up from the floor. The character is controlled by a deep neural network to perform the tasks. This work has been published to the notable conference on computer graphics SCA 2022.

Download the paper :

Activities:

- Learn deep neural network with a machine learning algorithm.

- Learn a physics-based character to get-up from floor using various styles.

Technologies:

- C++ and Python

- Vortex Engine

- Deep reinforcement learning (Use of PPO Algorithm)