[bugfix] EMA for the cascaded PID should start with the correct memory value #562
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
raul-openai
reviewed
Aug 14, 2020
| smooth_pos_setpoint = (smoothing_factor * ctrl_ema) + (1 - smoothing_factor) * d.ctrl[id] | ||
| # Apply Exponential Moving Average smoothing to the position setpoint, applying warmstart | ||
| # on the first iteration. | ||
| if d.time > 0.0: |
There was a problem hiding this comment.
Is this the best condition to check for? Ie: it happens at d.time == 0, but could it also happen under certain resets that clear data, but do not reset time?
Contributor
Author
There was a problem hiding this comment.
hmm I'm not sure if we reset the PID controller ever. are you aware of an action triggering that?
raul-openai
approved these changes
Aug 14, 2020
raul-openai
left a comment
raul-openai
left a comment
There was a problem hiding this comment.
Minor comment on condition, but otherwise 👍
arthurpetron
approved these changes
Aug 14, 2020
Comment on lines
+179
to
+183
| ctrl_ema = d.userdata[id * NUM_USER_DATA_PER_ACT + IDX_CAS_STORED_EMA_SMOOTH] | ||
| smoothing_factor = m.actuator_gainprm[id * NGAIN + IDX_CAS_EMA_SMOOTH] | ||
| smooth_pos_setpoint = (smoothing_factor * ctrl_ema) + (1 - smoothing_factor) * d.ctrl[id] | ||
| else: | ||
| smooth_pos_setpoint = d.ctrl[id] |
bayesian
reviewed
Aug 14, 2020
| smoothing_factor = m.actuator_gainprm[id * NGAIN + IDX_CAS_EMA_SMOOTH] | ||
| smooth_pos_setpoint = (smoothing_factor * ctrl_ema) + (1 - smoothing_factor) * d.ctrl[id] | ||
| else: | ||
| smooth_pos_setpoint = d.ctrl[id] |
Contributor
There was a problem hiding this comment.
shall we also set d.userdata[id * NUM_USER_DATA_PER_ACT + IDX_CAS_STORED_EMA_SMOOTH] to be 0 here, just to be safe.
Contributor
Author
There was a problem hiding this comment.
@bayesian the reassignment happens on the next line, is that needed?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
A previous bug introduced a behavior that started the EMA value at 0.0 that caused unintended qvel spikes due to drastic position change. This change allows the EMA to pass the first sample and initialize the EMA memory at the initial sample, avoiding any discontinuity issues.