Evolving Gaits for Damage Control in a Hexapod Robot

This repository contains all of the code necessary to replicate the simulated experiments from the paper Evolving Gaits for Damage Control in a Hexapod Robot.

Citing

If you use this work, please cite:

Mailer, C., Nitschke, G., and Raw, L. 2021. Evolving Gaits for Damage Control in a Hexapod Robot. In Proceedings of the Genetic and Evolutionary Computation Conference (pp. 146–153). Association for Computing Machinery.

BibTex:

@inproceedings{mailer2021,
  author = {Mailer, Christopher and Nitschke, Geoff and Raw, Leanne},
  title = {Evolving Gaits for Damage Control in a Hexapod Robot},
  year = {2021},
  publisher = {Association for Computing Machinery},
  doi = {10.1145/3449639.3459271},
  booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference},
  pages = {146–153},
  series = {GECCO '21} }

Python package dependencies

pip install -r requirements.txt

• numpy
• pybullet
• matplotlib
• sklearn
• mpi4py
• GPy
• scipy

Videos

Adaptation summary
Adapting to reality gap
Adapting to Failure Scenario S1
Adapting to Failure Scenario S2