MC acro: add separate params to configure yaw expo #9358

bkueng · 2018-04-23T12:49:49Z

So that it can be customized better according to personal preferences.

The limitation of 16 chars did not allow to use a better param name.

@hamishwillee This probably requires updates to the user-guide.

This matches the maximum rates for the attitude controller.

So that it can be customized better according to personal preferences. The limitation of 16 chars did not allow to use a better param name.

dagar · 2018-04-23T16:36:37Z

I realize why this currently makes sense, but could we briefly consider if things like manual input expo should be kept generic?

In the past we've discussed getting the manual handling completely out of the attitude and rate controllers for all vehicle types and moved into a new "stick mapper" module. At that point we could cleanly keep all of this input configuration generic with new features applying to all vehicle types.

I only mention this now to avoid possible parameter churn, unnecessary platform bloat (MC_ACRO_EXPO_Y/FW_ACRO_EXPO_Y/GND_ACRO_EXPO_Y), or fragmentation (feature gaps or nearly identical features implemented differently per vehicle now requiring different documentation).

hamishwillee · 2018-04-24T01:17:05Z

src/modules/mc_att_control/mc_att_control_params.c

+ * @decimal 2
+ * @group Multicopter Attitude Control
+ */
+PARAM_DEFINE_FLOAT(MC_ACRO_SUPEXPOY, 0.7f);


I would expect this to be MC_ACRO_SUPEXPO_Y for consistency and readability

As I wrote:

The limitation of 16 chars did not allow to use a better param name.

So yes your suggestion makes more sense, but exceeds the maximum length unfortunately.

hamishwillee · 2018-04-24T01:23:37Z

src/modules/mc_att_control/mc_att_control_params.c

@@ -395,8 +394,20 @@ PARAM_DEFINE_FLOAT(MC_ACRO_Y_MAX, 540.0f);
 PARAM_DEFINE_FLOAT(MC_ACRO_EXPO, 0.69f);


Should we rename MC_ACRO_EXPO_RP for consistency?

Renaming existing params is a pain, and I try to avoid it if at all possible. We can do it within the scope of a larger refactoring.

I think it should not be that big of a deal in this case. You can change the name if you want.

hamishwillee · 2018-04-24T01:29:13Z

src/modules/mc_att_control/mc_att_control_params.c

@@ -409,6 +420,20 @@ PARAM_DEFINE_FLOAT(MC_ACRO_EXPO, 0.69f);
 */
 PARAM_DEFINE_FLOAT(MC_ACRO_SUPEXPO, 0.7f);


MC_ACRO_SUPEXPO_RP ?

hamishwillee · 2018-04-24T01:30:38Z

@bkueng Assuming this goes ahead, the only place superexpo is mentioned is in the Acro mode.

As I understand it we now have three params that would need to be added.

MC_ACRO_EXPO (Roll, pitch) and MC_ACRO_EXPO_Y (Yaw). These define whether the sticks inputs are mapped linearly or as a cubic function.
MC_ACRO_SUPEXPOY and MC_ACRO_SUPEXPOY - IFF a cubic function, affects the curve to define how flat the response is - ie a larger value creates an effective "deadzone" of response in the centre of the joysticks.

Is that correct?

bkueng · 2018-04-24T07:13:38Z

@hamishwillee

MC_ACRO_SUPEXPOY and MC_ACRO_SUPEXPOY - IFF a cubic function, affects the curve to define how flat the response is - ie a larger value creates an effective "deadzone" of response in the centre of the joysticks.

I think it's easiest if you play with different param values and see how the curve changes: https://www.desmos.com/calculator/yty5kgurmc. g is the superexpo and f is the expo param, r is the maximum rate. Superexpo can be chosen entirely independent from expo.

@dagar

I only mention this now to avoid possible parameter churn, unnecessary platform bloat (MC_ACRO_EXPO_Y/FW_ACRO_EXPO_Y/GND_ACRO_EXPO_Y), or fragmentation (feature gaps or nearly identical features implemented differently per vehicle now requiring different documentation).

I agree, we should not have one expo param per type. But what you definitely want is different settings for different flight modes. I suggest the following steps moving forward:

Merge this as is if everyone agrees
when introducing expo to fixed-wing acro (which is planned), choose names that are vehicle-unspecific
do the refactoring of MC acro params to use the generic ones
do the internal refactoring & stick mapper and more renaming if needed.

hamishwillee · 2018-04-26T06:24:31Z

MC_ACRO_SUPEXPOY and MC_ACRO_SUPEXPOY - IFF a cubic function, affects the curve to define how flat the response is - ie a larger value creates an effective "deadzone" of response in the centre of the joysticks.

I think it's easiest if you play with different param values and see how the curve changes: https://www.desmos.com/calculator/yty5kgurmc. g is the superexpo and f is the expo param, r is the maximum rate. Superexpo can be chosen entirely independent from expo.

That helps A LOT thanks.

While strictly accurate, it is hard to understand the relationships from the descriptions in the parameters. I'm not even sure that someone would necessarily understand that these parameters help provide the mapping between stick position and setpoint.

While we could try an explain all this in the parameters, I suggest we update the Acro mode docs to explain this and use your link to demonstrate it. We might then link to the new guide from the parameter documentation?

MC_ACRO_EXPO(FLOAT) 
- Acro Expo factor applied to input of pitch and roll axes (see also [Link to Acro mode input scaling])
- Comment: 0 Purely linear input curve 1 Purely cubic input curve

MaEtUgR · 2018-04-26T09:46:43Z

Merge this as is if everyone agrees

when introducing expo to fixed-wing acro (which is planned), choose names that are vehicle-unspecific

do the refactoring of MC acro params to use the generic ones

do the internal refactoring & stick mapper and more renaming if needed.

Exactly what I wanted to suggest.

@hamishwillee See #8036 for the original implementation pr.

ie a larger value creates an effective "deadzone" of response in the centre of the joysticks.

Not exactly, the goal is to have a high turn rate with maximum stick input but to have a zone of lower sensitivity for small corrections close to the stick center. A deadzone is not lower sensitivity, it's a zone where there's zero action to the stick to compensate for hardware tolerances and calibration. According to my opinion you don't want to have a deadzone in acro mode because results in a gap for your anyways necessary contiunous control input. A deadzone is desirable e.g. in position hold such that when you let go of the stick it does not drift at all because of your stick inaccuracy.

MaEtUgR

I thought about it when implementing the rc curves but assumed I'm only increasing parameter number without someone actually using a different yaw value. And it's easy to add like you did right now. lgtm

MaEtUgR · 2018-04-26T11:02:54Z

src/modules/mc_att_control/mc_att_control_params.c

@@ -347,7 +347,7 @@ PARAM_DEFINE_FLOAT(MC_YAWRAUTO_MAX, 45.0f);
 *
 * @unit deg/s
 * @min 0.0
- * @max 1000.0
+ * @max 1800.0


Did you do experiements about the overshoot not accidentally breaching the 2000deg/s gyro limit barrier? I don't have experience with >1000 deg/s rates.

No, not that high, but I did go over 1000. Would be interesting to know indeed.
Looks like others are actually going to the 2000 limit: betaflight/betaflight-configurator#997

bkueng · 2018-04-26T12:16:56Z

@hamishwillee I agree, it's very hard to explain it with a textual description only. We can add the link to the param documentation, but I'm not sure if it helps, because I don't think QGC will show it as clickable link.
The fancy solution would be that QGC shows the curve on a stick mapping setup page.

hamishwillee · 2018-04-27T06:28:54Z

@bkueng @MaEtUgR That is very helpful thank you.

I have created a stick mapping section in Acro mode docs for old/current implementation - that can be extended when this is accepted:
- PR
- rendered
Suggest we consider an update to the parameter descriptions to something like below. They aren't perfect but they do make it more clear that this is Acro mode, that expo stands for exponential, that it is for tuning the input curve, and that the supexpo (I just see it isn't "super" but "sup" grrr) is related to the expo.

MC_ACRO_EXPO - Acro mode "exponential" factor for tuning input curve shape (applied to input of all axis: roll, pitch, yaw)
MC_ACRO_SUPEXPO - Acro mode "superexpo" factor for refining input curve shape tuned using MC_ACRO_EXPO (applied to input of all axes: roll, pitch, yaw).

MaEtUgR · 2018-05-02T08:02:25Z

@hamishwillee That's very nice documentation, if desired I could provide a calculator to get the maximal angular rate, expo and superexpo parameters from the wide spread not independent parametrized rates configuration of betaflight and all the related https://oscarliang.com/rc-roll-pitch-yaw-rate-cleanflight/. Such that users who have some experience with this configuration can convert their favorite settings... Probably this would even make sense to be embedded in the UI.

hamishwillee · 2018-05-03T04:18:43Z

@MaEtUgR Well, it is certainly better than what was there before (nothing) :-). I like betafight's explanation too, though I have tried to be more terse and let the diagrams do the talking. Do you think it is worth/OK to provide that link as "further information"?

I don't understand exactly what you are proposing above (ie I don't know what "wide spread" means in " provide a calculator to get the maximal angular rate, expo and superexpo parameters from the wide spread"

Are you suggesting you can create a tool to get a user's settings in CleanFlight and convert that into our parameters?

If not, can you explain what you are proposing, who would use it, and where would it live? Sorry if I am being dumb.

MaEtUgR · 2018-06-04T16:01:54Z

@hamishwillee Sry for reaction time. Nothing to worry about your explanation, I'm not suggesting to change it.

What I was offering is a small conversion tool to use on the website that converts the parameters most race firmwares which indirectly fork from cleanflight, betaflight and all these to ours. They use RC Rate, RC Expo, Super Rate which all depend on each other and are therefore hard to tune. I could write up a small script for the calculations back and forth.

EDIT: I created an issue for that: #9591

bkueng added 2 commits April 23, 2018 14:38

mc_att_control_params: increase max acro rates to 1800

f63739c

This matches the maximum rates for the attitude controller.

MC acro: add separate params to configure yaw expo

b68e6b1

So that it can be customized better according to personal preferences. The limitation of 16 chars did not allow to use a better param name.

bkueng requested a review from MaEtUgR April 23, 2018 12:49

hamishwillee reviewed Apr 24, 2018

View reviewed changes

MaEtUgR approved these changes Apr 26, 2018

View reviewed changes

bkueng merged commit f2cee99 into master Apr 26, 2018

bkueng deleted the acro_yaw_expo branch April 26, 2018 12:17

bkueng mentioned this pull request Apr 27, 2018

MC acro: update param descriptions #9374

Merged

MaEtUgR mentioned this pull request Jun 4, 2018

Convert racer joystick rates to PX4 acro rates #9591

Open

mrpollo mentioned this pull request Jun 5, 2018

v1.8 Changelog scratchpad #9593

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

MC acro: add separate params to configure yaw expo #9358

MC acro: add separate params to configure yaw expo #9358

bkueng commented Apr 23, 2018 •

edited by AuterionWrikeBot

Loading

dagar commented Apr 23, 2018

hamishwillee Apr 24, 2018

bkueng Apr 24, 2018

hamishwillee Apr 24, 2018

bkueng Apr 24, 2018

MaEtUgR Apr 26, 2018

hamishwillee Apr 24, 2018

hamishwillee commented Apr 24, 2018

bkueng commented Apr 24, 2018

hamishwillee commented Apr 26, 2018

MaEtUgR commented Apr 26, 2018 •

edited

Loading

MaEtUgR left a comment

MaEtUgR Apr 26, 2018 •

edited

Loading

bkueng Apr 26, 2018

bkueng commented Apr 26, 2018

hamishwillee commented Apr 27, 2018 •

edited

Loading

MaEtUgR commented May 2, 2018 •

edited

Loading

hamishwillee commented May 3, 2018

MaEtUgR commented Jun 4, 2018 •

edited

Loading

		@@ -395,8 +394,20 @@ PARAM_DEFINE_FLOAT(MC_ACRO_Y_MAX, 540.0f);
		PARAM_DEFINE_FLOAT(MC_ACRO_EXPO, 0.69f);

		@@ -409,6 +420,20 @@ PARAM_DEFINE_FLOAT(MC_ACRO_EXPO, 0.69f);
		*/
		PARAM_DEFINE_FLOAT(MC_ACRO_SUPEXPO, 0.7f);

MC acro: add separate params to configure yaw expo #9358

MC acro: add separate params to configure yaw expo #9358

Conversation

bkueng commented Apr 23, 2018 • edited by AuterionWrikeBot Loading

dagar commented Apr 23, 2018

hamishwillee Apr 24, 2018

Choose a reason for hiding this comment

bkueng Apr 24, 2018

Choose a reason for hiding this comment

hamishwillee Apr 24, 2018

Choose a reason for hiding this comment

bkueng Apr 24, 2018

Choose a reason for hiding this comment

MaEtUgR Apr 26, 2018

Choose a reason for hiding this comment

hamishwillee Apr 24, 2018

Choose a reason for hiding this comment

hamishwillee commented Apr 24, 2018

bkueng commented Apr 24, 2018

hamishwillee commented Apr 26, 2018

MaEtUgR commented Apr 26, 2018 • edited Loading

MaEtUgR left a comment

Choose a reason for hiding this comment

MaEtUgR Apr 26, 2018 • edited Loading

Choose a reason for hiding this comment

bkueng Apr 26, 2018

Choose a reason for hiding this comment

bkueng commented Apr 26, 2018

hamishwillee commented Apr 27, 2018 • edited Loading

MaEtUgR commented May 2, 2018 • edited Loading

hamishwillee commented May 3, 2018

MaEtUgR commented Jun 4, 2018 • edited Loading

bkueng commented Apr 23, 2018 •

edited by AuterionWrikeBot

Loading

MaEtUgR commented Apr 26, 2018 •

edited

Loading

MaEtUgR Apr 26, 2018 •

edited

Loading

hamishwillee commented Apr 27, 2018 •

edited

Loading

MaEtUgR commented May 2, 2018 •

edited

Loading

MaEtUgR commented Jun 4, 2018 •

edited

Loading