Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

CHEVROLET BOLT EUV 2022: Add a simple neural feed forward #31266

Merged
merged 4 commits into from Feb 1, 2024

Conversation

nuwandavek
Copy link
Contributor

Note: Moving PR from #31255 for simplicity

  • We already customize lateral feed-forward for platforms in various ways (Static torque control, torqued, non-linear response etc.).
  • This PR introduces a neural feed forward for the CHEVROLET BOLT EUV 2022 platform.
  • This is a small 4-layer MLP (262 parameters) that takes in the lateral_acceleration, roll_compensation, vego and aego
  • The model was then evaluated in the MLControlsSimulator, allowing easy comparison between the various feed-forward strategies for the Bolt:

  • Examples live in the car:

@github-actions github-actions bot added the car vehicle-specific label Feb 1, 2024
@nuwandavek nuwandavek added controls controls/planner related gm labels Feb 1, 2024
@nuwandavek nuwandavek merged commit 6196256 into master Feb 1, 2024
26 of 27 checks passed
@nuwandavek nuwandavek deleted the bolt-neural-simple branch February 1, 2024 20:30
@simontheflutist
Copy link

Did this commit break master-ci? The loader is there, but selfdrive/car/torque_data/neural_ff_weights.json is missing as of this morning's 8d799ec.

simontheflutist added a commit to simontheflutist/openpilot that referenced this pull request Feb 2, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
car vehicle-specific controls controls/planner related gm
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants