This project explores evolutionary algorithms to optimize soft robots in EvoGym. It covers three core areas:
-
Evolving Structure
Optimization of robot morphology (voxel layout) using evolutionary strategies to improve task performance. -
Evolving Controller
Evolution of neural controllers for a fixed robot structure, focusing on locomotion and task-solving ability. -
Evolving Structure and Controller
Co-evolution of both the robot's physical structure and its controller using cooperative optimization methods.
- Python 3.10
- EvoGym 2.0.0
- NumPy 1.23.5
- Torch 2.5.1
- Gymnasium 2.0.0
- Imageio 2.36.1
For more info on EvoGym, visit evolutiongym.github.