OCCRA
Go to Post FIRST will continue searching for rules and procedures to keep us from hurting ourselves. However, none of this will substitute for judicious application of common sense. - Richard Wallace [more]
Home
Go Back   Chief Delphi > ChiefDelphi.com Website > Extra Discussion
CD-Media  
portal register members calendar search Today's Posts Mark Forums Read FAQ rules

 
Reply
Thread Tools Rate Thread Display Modes
  #16   Spotlight this post!  
Unread 08-20-2018, 12:52 PM
Andrew Schreiber Andrew Schreiber is offline
Petitioning the UN to ban Lythgoe
no team
Team Role: Engineer
 
Join Date: Jan 2005
Rookie Year: 2000
Location: Somehow, the entire east coast...
Posts: 4,952
Andrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Nick_Coussens View Post
Did you guys have any issues sending the CAN through the slip ring? We have discussed it in the past but felt this would be a risk point. What slip ring did you use?
There were no CAN errors in the logs from the slip rings. You'd have to ask John for the particular model, I don't have it handy.
__________________




.
Reply With Quote
  #17   Spotlight this post!  
Unread 08-20-2018, 12:56 PM
marshall's Avatar
marshall marshall is online now
"Who's Marshall?"
FRC #0900 (The Zebracorns)
Team Role: Mentor
 
Join Date: Jan 2012
Rookie Year: 2003
Location: North Carolina
Posts: 3,086
marshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Nick_Coussens View Post
Did you guys have any issues sending the CAN through the slip ring? We have discussed it in the past but felt this would be a risk point. What slip ring did you use?
It's on our list to test this more scientifically. We've got a couple of the rings and need to invest some time in a test rig. It seems to work well for the teams using it but I'm skeptical until it's been tested and documented. It also necessitates a topology change to make it work.
__________________
"La mejor salsa del mundo es la hambre" - Miguel de Cervantes
"The future is unwritten" - Joe Strummer
"Simplify, then add lightness" - Colin Chapman
Clarke's Laws
Reply With Quote
  #18   Spotlight this post!  
Unread 08-20-2018, 01:54 PM
nickbrickmaster nickbrickmaster is offline
Registered User
AKA: Nick Schatz
no team ('Snow Problem, 3184 Alum)
Team Role: Programmer
 
Join Date: Jan 2015
Rookie Year: 2014
Location: Eagan MN
Posts: 478
nickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Chris is me View Post
This isn't a "psuedo differential" - this is 100% a differential swerve, it just uses the ground / friction to transmit that differential force.

I've toyed with this concept a little bit as well, as it avoids the bevel gears altogether which is nice. The 4143 module that you linked to is the other one I've seen. So far it's been really challenging to package it better than bevel gear differential swerves - and you can sorta see that here.
The packaging is definitely what I dislike the most about this module. I've been running over this in my head for the past few days, and the only thing I've come up with that would remove a significant amount of height is to get rid of the unnecessarily long belt run. However this requires significantly changing the design, either moving everything below the bearing (which runs into significant cant forces) or putting the reduction into the plane of the bearing, requiring a significantly larger bearing.

If I have the time over the next few days I may explore the latter option a bit with a derivative of this design.

Unless the packaging gets a lot, lot better I don't see any reason to run a traction-differential swerve, since gear based designs package much better and also can be built with no more beveling than a traditional swerve (@anand)
__________________
This is a postmodern signature.
Reply With Quote
  #19   Spotlight this post!  
Unread 08-20-2018, 09:16 PM
asid61's Avatar
asid61 asid61 is offline
Design Simple
AKA: Anand Rajamani
FRC #1072 (Harker Robotics)
Team Role: Mentor
 
Join Date: Jan 2014
Rookie Year: 2013
Location: Cupertino, CA
Posts: 3,125
asid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Chris is me View Post
This isn't a "psuedo differential" - this is 100% a differential swerve, it just uses the ground / friction to transmit that differential force.

I've toyed with this concept a little bit as well, as it avoids the bevel gears altogether which is nice. The 4143 module that you linked to is the other one I've seen. So far it's been really challenging to package it better than bevel gear differential swerves - and you can sorta see that here.
I thought the packaging was bad until I realized that you cut out the pivot gearbox altogether with this design- it just becomes a 16-style module with an extra 775.
At the very least, modules like 2767's are much larger by comparison.
__________________
Team 1072 2017-present
Team 299 2017
Team 115 2013-2016 (student)

2018 Davis Finalists (w/ 6474 and 3880), 2018 Roebling Winners (w/ 3476, 1323, and 1778)

Reply With Quote
  #20   Spotlight this post!  
Unread 08-21-2018, 08:48 AM
Chris is me's Avatar
Chris is me Chris is me is offline
on a break
AKA: Pinecone
no team (formerly FRC 3929, 228, 2791)
Team Role: Mentor
 
Join Date: Dec 2008
Rookie Year: 2006
Location: Chicago, IL
Posts: 8,573
Chris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond reputeChris is me has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by asid61 View Post
I thought the packaging was bad until I realized that you cut out the pivot gearbox altogether with this design- it just becomes a 16-style module with an extra 775.
At the very least, modules like 2767's are much larger by comparison.
Just depends on if you need the height vs the width more. A lot of games I'd prefer to take up more space in and under the chassis plane than above it, but I suppose just as many games with tighter packaging down low exist. Just another design factor to think about.

I don't want to suggest making the module ring wider, but I feel like if you did, then you could tuck the motors within the module so that they don't have to clear the gears as much? The gears on each side don't HAVE to be coaxial... Just something to think about.
__________________
Sort-of looking for a new team in or near Chicago; PM me if you have a lead.
Mentor: 3929 (2018)
Mentor / Drive Coach: 228 (2016-17) - 2016 RIDE Winner &
Consultant: 2170 (2017)
College Mentor: 2791 (2010-15) Build Photos - 2013 WPI Finalists, 2012 BAE Finalists
Student: 1714 (2009) - 2009 MN 10K Lakes Winners
Reply With Quote
  #21   Spotlight this post!  
Unread 08-25-2018, 02:35 AM
jrod jrod is offline
Registered User
AKA: Jackson McLaughlin
FRC #5818 (Riviera Robotics)
Team Role: Student
 
Join Date: Jun 2017
Rookie Year: 2017
Location: Santa Barbara, CA
Posts: 9
jrod is a glorious beacon of lightjrod is a glorious beacon of lightjrod is a glorious beacon of lightjrod is a glorious beacon of lightjrod is a glorious beacon of lightjrod is a glorious beacon of light
Re: pic: Pseudo-Differential Swerve

We were experimenting with a swerve like this a bit before the 2017 build season. We had encoders on each wheel and one for the module azimuth. The hardware worked fine and we didn't have issues with CAN through the slip ring or with the star topology. Even on flat carpet, we had controllability issues with wheels losing traction. We had a PID azimuth loop with an output that sent opposite power to each wheel, summed with a feed-forward for forward/backwards drive. For some brief moments, it moved in control, but anytime one wheel would lose traction, the module would spin. Once this started to happen, the module would be out of control until it got dragged to an area where the wheels had equal traction. With only 2 modules on the test base, this usually took too long and threw the control of the whole robot off. After seeing these issues, we mostly abandoned the design.

With 4 modules, the control would probably be improved some due to higher probability of at least 2 modules in control. Another thing that could help would be running the loops on the Talon for higher frequency. We couldn't do this when we first tried it because it requires some new talon features discussed here to have an azimuth loop running at the same time as the feedforward for the forwards/backwards drive of the module. Still, we don't plan on trying a swerve like this again, since not being able to control a module that loses traction on one wheel seems like too big of an issue.

Imgur Album
Reply With Quote
  #22   Spotlight this post!  
Unread 08-27-2018, 08:40 AM
Skyehawk's Avatar
Skyehawk Skyehawk is offline
Nuts N' Bolts
AKA: Skye Leake
FRC #2767 (Stryke Force, Thunder Robotics, Team F.U.N.)
Team Role: Mentor
 
Join Date: Nov 2012
Rookie Year: 2011
Location: Kalamazoo, MI
Posts: 702
Skyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Correct me if I'm wrong, (who am I kidding, this is CD, you WILL let me know if I'm wrong ).

Isn't the minimum number of sensors on a differential swerve utilizing the ground to induce the azimuth change moment 3? One for each "side" of the drivetrain and one for absolute azimuth heading. More sensors = more $$$ & tonnage.

On a "conventional" differential swerve where the azimuth change moment is handled by the gearing you can get away with 2 sensors, one on a motor and one for the absolute azimuth. (You can simply match the sensored motor to hold the azimuth to the angular velocity/position desired). This is possible because the whole lot is indexed together by the gearing. Whereas the OP design is prone to losing the index position through wheel slip, therefore it needs the position of both sides of the drivetrain to calculate distance.

I also propose we call this design of diff swerve: "Drive Surface Indexed Differential Swerve". Vs. conventional diff swerve being called: "Internally Indexed Differential Swerve".
__________________
My time in FIRST has made me a better person, frankly I don't know where I'd be without it.
  • A small mountain of hardware, good memories & life lessons with 876: 2011 - 2018
  • Inevitable good times with 2767 : 2018 - ????

Last edited by Skyehawk : 08-27-2018 at 08:47 AM. Reason: nomenclature
Reply With Quote
  #23   Spotlight this post!  
Unread 08-27-2018, 09:10 AM
Andrew Schreiber Andrew Schreiber is offline
Petitioning the UN to ban Lythgoe
no team
Team Role: Engineer
 
Join Date: Jan 2005
Rookie Year: 2000
Location: Somehow, the entire east coast...
Posts: 4,952
Andrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Skyehawk View Post
Correct me if I'm wrong, (who am I kidding, this is CD, you WILL let me know if I'm wrong ).

Isn't the minimum number of sensors on a differential swerve utilizing the ground to induce the azimuth change moment 3? One for each "side" of the drivetrain and one for absolute azimuth heading. More sensors = more $$$ & tonnage.

On a "conventional" differential swerve where the azimuth change moment is handled by the gearing you can get away with 2 sensors, one on a motor and one for the absolute azimuth. (You can simply match the sensored motor to hold the azimuth to the angular velocity/position desired). This is possible because the whole lot is indexed together by the gearing. Whereas the OP design is prone to losing the index position through wheel slip, therefore it needs the position of both sides of the drivetrain to calculate distance.

I also propose we call this design of diff swerve: "Drive Surface Indexed Differential Swerve". Vs. conventional diff swerve being called: "Internally Indexed Differential Swerve".
I mean, that really depends on what your sensors are. You COULD tell everything you need to know using two absolute encoders 1:1 with the inputs to a differential swerve since you'd know the absolute angle of both inputs. It's likely not practical but it's possible.

Or, perhaps you want to rely on the fact that the rotation speed is driven by the relationship between the inputs and thus you could measure one input and the azimuth and compute the other input. That's likely the easiest to do.

Of course, my mental model is the Aren Hill style compact module where I can effectively instrument things fairly cleaning. I'm still trying to wrap my head around this style module, so I'm not 100% certain where these sensors would need to go. Realistically, yes this needs 3 sensors for easiest operation. Though one of those sensors could be as simple as a hall effect switch for zeroing.

So, I'm not saying you're wrong, just that you're maybe not sufficiently bonkers?
__________________




.
Reply With Quote
  #24   Spotlight this post!  
Unread 08-27-2018, 09:32 AM
Skyehawk's Avatar
Skyehawk Skyehawk is offline
Nuts N' Bolts
AKA: Skye Leake
FRC #2767 (Stryke Force, Thunder Robotics, Team F.U.N.)
Team Role: Mentor
 
Join Date: Nov 2012
Rookie Year: 2011
Location: Kalamazoo, MI
Posts: 702
Skyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Andrew Schreiber View Post
I mean, that really depends on what your sensors are. You COULD tell everything you need to know using two absolute encoders 1:1 with the inputs to a differential swerve since you'd know the absolute angle of both inputs. It's likely not practical but it's possible.

Or, perhaps you want to rely on the fact that the rotation speed is driven by the relationship between the inputs and thus you could measure one input and the azimuth and compute the other input. That's likely the easiest to do.

Of course, my mental model is the Aren Hill style compact module where I can effectively instrument things fairly cleaning. I'm still trying to wrap my head around this style module, so I'm not 100% certain where these sensors would need to go. Realistically, yes this needs 3 sensors for easiest operation. Though one of those sensors could be as simple as a hall effect switch for zeroing.

So, I'm not saying you're wrong, just that you're maybe not sufficiently bonkers?
Consider this:
The module is traveling along in a straight line, all of a sudden an opposing robot hits it from the side causing the non-sensored wheel to remain on the ground while the sensored wheel is in the air. Less rolling friction on the sensored wheel (in the air) makes it go faster, the non-sensored side of the module continues as before. Since the sensored side of the module is our only count of distance we now assume the robot has traveled further than it has in reality.

This problem can be simplified to that of a skidsteer robot with an encoder on one side of the drivetrain and a gyro for absolute position:
A robot is traveling along in a straight line all of a sudden an opposing robot hits it from the side causing the non-sensored side of the drivetrain to remain on the ground while the sensored side of the drivetrain is in the air. Less rolling friction on the sensored side (in the air) makes it go faster, the non-sensored side of the drivetrain continues as before. Since the sensored side of the drivetrain is our only count of distance we now assume the robot has traveled further than it has in reality.

Now, if we put sensors on both sides of the module/drivetrain we at least know of the discrepancy in distance.
__________________
My time in FIRST has made me a better person, frankly I don't know where I'd be without it.
  • A small mountain of hardware, good memories & life lessons with 876: 2011 - 2018
  • Inevitable good times with 2767 : 2018 - ????
Reply With Quote
  #25   Spotlight this post!  
Unread 08-27-2018, 10:47 AM
Andrew Schreiber Andrew Schreiber is offline
Petitioning the UN to ban Lythgoe
no team
Team Role: Engineer
 
Join Date: Jan 2005
Rookie Year: 2000
Location: Somehow, the entire east coast...
Posts: 4,952
Andrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Skyehawk View Post
Consider this:
The module is traveling along in a straight line, all of a sudden an opposing robot hits it from the side causing the non-sensored wheel to remain on the ground while the sensored wheel is in the air. Less rolling friction on the sensored wheel (in the air) makes it go faster, the non-sensored side of the module continues as before. Since the sensored side of the module is our only count of distance we now assume the robot has traveled further than it has in reality.

This problem can be simplified to that of a skidsteer robot with an encoder on one side of the drivetrain and a gyro for absolute position:
A robot is traveling along in a straight line all of a sudden an opposing robot hits it from the side causing the non-sensored side of the drivetrain to remain on the ground while the sensored side of the drivetrain is in the air. Less rolling friction on the sensored side (in the air) makes it go faster, the non-sensored side of the drivetrain continues as before. Since the sensored side of the drivetrain is our only count of distance we now assume the robot has traveled further than it has in reality.

Now, if we put sensors on both sides of the module/drivetrain we at least know of the discrepancy in distance.
Wouldn't one side no longer being in contact with the ground result in a torque about the azimuth axis rotating wheel which would show up in the azimuth sensor?
__________________




.
Reply With Quote
  #26   Spotlight this post!  
Unread 08-27-2018, 11:36 AM
Skyehawk's Avatar
Skyehawk Skyehawk is offline
Nuts N' Bolts
AKA: Skye Leake
FRC #2767 (Stryke Force, Thunder Robotics, Team F.U.N.)
Team Role: Mentor
 
Join Date: Nov 2012
Rookie Year: 2011
Location: Kalamazoo, MI
Posts: 702
Skyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Andrew Schreiber View Post
Wouldn't one side no longer being in contact with the ground result in a torque about the azimuth axis rotating wheel which would show up in the azimuth sensor?
Not necessarily (although I think it would in most cases), putting aside outside influence holding things straight (i.e. another robot in a collision), albeit we are getting into a fairly specific case here...

An impact causing the sensored wheel to leave the ground orthogonal to the direction of motion (side collision) will bring the contact point of the non-sensored wheel closer to the intersection of the azimuth axis & the ground, this would reduce the magnitude of the azimuth change moment. This problem only gets worse with narrow packaging of wheels* and larger diameter wheels. I am not entirely sure if this is a stable config (I think it is if it is overcentered), but it doesn't have to be if there is not enough time for the azimuth to change while the sensored wheel is off the ground (see poorly put together exaggerated example)

https://i.imgur.com/5qZqctJ.png


*more stable under standard use, whereas wider packaging can supply a larger & more accurate azimuth change

I am thinking the solution to all this is to place the azimuth bearing low on the module, at or below the wheel axle.
__________________
My time in FIRST has made me a better person, frankly I don't know where I'd be without it.
  • A small mountain of hardware, good memories & life lessons with 876: 2011 - 2018
  • Inevitable good times with 2767 : 2018 - ????

Last edited by Skyehawk : 08-27-2018 at 11:41 AM. Reason: clarification
Reply With Quote
  #27   Spotlight this post!  
Unread 08-27-2018, 11:54 AM
Andrew Schreiber Andrew Schreiber is offline
Petitioning the UN to ban Lythgoe
no team
Team Role: Engineer
 
Join Date: Jan 2005
Rookie Year: 2000
Location: Somehow, the entire east coast...
Posts: 4,952
Andrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond reputeAndrew Schreiber has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

I always assume that robot odometery will be wildly off following any sort of collision anyway so yeah, VERY specific case here.

Putting your azimuth bearing low is beneficial for other reasons too as I'm certain you know. Mostly dealing with minimizing load on your modules should the bot be tipped onto the side of the modules. If it helps maintain localization information in the case of an impact that's just icing imho.
__________________




.
Reply With Quote
  #28   Spotlight this post!  
Unread 08-27-2018, 12:11 PM
AriMB's Avatar
AriMB AriMB is online now
The Philadelphian emigrant
AKA: Ari Meles-Braverman
FRC #5987 (Galaxia)
Team Role: Mentor
 
Join Date: Mar 2015
Rookie Year: 2012
Location: Haifa, Israel
Posts: 1,832
AriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond reputeAriMB has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Andrew Schreiber View Post
I always assume that robot odometery will be wildly off following any sort of collision anyway so yeah, VERY specific case here.

Putting your azimuth bearing low is beneficial for other reasons too as I'm certain you know. Mostly dealing with minimizing load on your modules should the bot be tipped onto the side of the modules. If it helps maintain localization information in the case of an impact that's just icing imho.
It's not just after a collision though. Anything that would cause one wheel to slip and not the other would cause the same problem, even if the wheel isn't completely raised off the ground. This could happen in autonomous by driving over a small bump* or a slippery/sticky piece of field tape**, for example. If the wheel that's slipping is the one you're measuring, you may know that the wheel slipped (because the module turned unexpectedly), but you don't know how much distance you lost from slipping.


* flat fields often have unexpected 1/4" bumps (see 2017)
** FTAs do their best to keep all the tape lines intact and smooth, but you know as well as I that they get messed up by the end of a long competition
__________________
Studying MechE at the Technion - Israel Institute of Technology
2017-present: FIRST Israel CSA/FTAA
2017-present: FRC 5987 Technical Mentor 18isr2 18isr4 18isrcmp 18carv
2012-2016: FRC 423 Member 15njtab 15padre 16paphi
Reply With Quote
  #29   Spotlight this post!  
Unread 08-27-2018, 12:18 PM
Skyehawk's Avatar
Skyehawk Skyehawk is offline
Nuts N' Bolts
AKA: Skye Leake
FRC #2767 (Stryke Force, Thunder Robotics, Team F.U.N.)
Team Role: Mentor
 
Join Date: Nov 2012
Rookie Year: 2011
Location: Kalamazoo, MI
Posts: 702
Skyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond reputeSkyehawk has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by AriMB View Post
<clip> for example. If the wheel that's slipping is the one you're measuring, you may know that the wheel slipped (because the module turned unexpectedly), but you don't know how much distance you lost from slipping.

</clip>
This, although the module may not always turn due to a variety of reasons that I am stull thinking about.
__________________
My time in FIRST has made me a better person, frankly I don't know where I'd be without it.
  • A small mountain of hardware, good memories & life lessons with 876: 2011 - 2018
  • Inevitable good times with 2767 : 2018 - ????

Last edited by Skyehawk : 08-27-2018 at 12:22 PM.
Reply With Quote
  #30   Spotlight this post!  
Unread 08-27-2018, 01:21 PM
nickbrickmaster nickbrickmaster is offline
Registered User
AKA: Nick Schatz
no team ('Snow Problem, 3184 Alum)
Team Role: Programmer
 
Join Date: Jan 2015
Rookie Year: 2014
Location: Eagan MN
Posts: 478
nickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond reputenickbrickmaster has a reputation beyond repute
Re: pic: Pseudo-Differential Swerve

Quote:
Originally Posted by Skyehawk View Post
This, although the module may not always turn due to a variety of reasons that I am stull thinking about.
You'd still have the module azimuth (with the 3-sensor setup I have here), but you'd lose odometry. It's equivalent to a tank drive slipping on one side- you'll lose counts.
Probably because this is so easy to slip (less traction per wheel than a standard tank drive or swerve) you'd want some sort of follower wheels for odometry.

The real problem with a wheel slipping is that the model becomes invalidated. Unless you massively overcomplicate the model, you would want to assume it's like a normal differential swerve, where the difference of speeds produces rotation. As soon as one wheel loses traction, this is invalid. At worst, you'd lose module control entirely until you gain traction again. At best (probably closer to the truth, but depending on relative gains) you'd have slightly dampened azimuth control but be unable to drive in anything other than a very tight circle.

You might be able compensate for this by measuring speed, voltage, and current to estimate torque output and modeling wheel rotation and and speed as a function of these, but that's a massively overcomplicated model for marginal gain.
__________________
This is a postmodern signature.
Reply With Quote
Reply


Thread Tools
Display Modes Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 11:14 AM.

The Chief Delphi Forums are sponsored by Innovation First International, Inc.


Powered by vBulletin®
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi