commaai / commaai/opendbc

2024 Kia Niro steering humming during override in sharp turns

Open
#3,685 0 comments 0 reactions 0 assignees View on GitHub
bug car
Dominant language
Python
Stars
3.4k
Forks
2.3k
Avg merge
3d 2h
Merged PRs (30d)
38

Description

### Describe the car bug

Which device are you using?
comma four

Describe the bug

When taking a sharp turn under brake (such as taking a right turn onto a street or into a parking spot) while comma four is engaged in lateral control, I'll get a vibration that feels like the comma is trying to do contradictory inputs. It reminds me of a PWM vibration in a DC motor, like it's making micro adjustments to the steering as I turn. It doesn't impact control of the vehicle, but I dread it because it feels like I'm grinding gears. This does not occur with lat disengaged, and is not the collision warning system that vibrates the steering wheel, that is a different feeling that I've made a point of differentiating during my own troubleshooting.

I expect to feel continous resistance during turns, same as when I make steering inputs at other angles. Turning down the friction value from 0.9 to 0.8 in steering in Sunnylink seems to alleviate the worst of it, but I find the the car wanders in its lane more so that doesn't feel like the solution either. I'm running North Nevada 2 right now and there is a strong correlation between strong lane keeping and this vibration - so either I get sloppy lane keeping or I get grinding turns.

I'm not sure how to reproduce but here is a related issue that is still open and I believe is relevant to my case: https://github.com/commaai/opendbc/issues/1112

In the route I've shared, the vibration occurs during the turn around 8:13:17.

Route is from an older version but I still see this happening on the latest. I can find newer routes but I couldn't easily find one where it happened so I'm providing one I used for a bug report on the sunnypilot community.

### Provide a route where the issue occurs

0d4257d1c6741384/000001a8--af30338338/9

### openpilot version

Sunnypilot 2026.04.12-4387

### Additional info

_No response_

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reviewing related opendbc issue #1112 and the provided route, focusing on the turn around 8:13:17. Compare the reported behavior with the latest version rather than the older route version, and use the friction-value and lateral-control details as investigation context. Done means identifying the cause of the steering vibration and confirming a fix without weakening lane keeping.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
embedded-iot
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.