Gripper shoes for XGO

[Steve " 'dillo" Okay]

The LuwuDynamics XGO-series of quadruped robots have hard nylon legs that are good for a product where you don't know what sort of surface somebody might use it on. In practice, they kind of suck for traction indoors. They're slippery and the robot has a shuffling,sliding kind of gait vs. a walk.

I tried gasket foam and trayliner mat, but those all either wore easily or didn't grip much(if any). So I did some CAD in OnShape, fired up the printer w/ a spool of TPU and a few hours later, little rubber dog-shoes!

It looks like this:

and works like this:
https://youtu.be/1I5Dhr9_iXA

[Chris Albertson]

I did some experiments as I also have a hardwood floor.

The most expspedient fix was “friction tape”. It is mmade by 3M and sold at Home Depot and Amazon.  It is used by electricians under that more widely known vinyl electrical tape.    The tape is made of cloth but with rubber infused into it andonly slightly sticky on one side.     I place ed multiple layers on the feet then secured it with a few wraps of the vinyl tale around the “ankles”.   It does not leave black marks on the floor.

I also tried putting a very thin layer of foam sheet under the tape.   It might have helped but it is not durable.

A better solution is to change the foot design.  The classic design is to buy “squash ball” and cut them in half and use those for feet.  The MIT Cheetah used these and many have copied the idea.  Solid rubber balls are also the perfect shape for feet.

My solution was to print feet in TPU.  You can adjust th softness by buying the “correct” TPU but also by varying the shell thickness and infill percentage.

My next robot will have pressure sensors on the feet and I want it to work outdoors.   

Friction tape was a simple and quick fix

Oh and it is is “shuffling” look at the software.   Likely the kinematics is wrong.  Maybe not taking into concideration compliance in the parts and using theoretical as-designed dimension rather then as-standing dimensions.

On Jul 9, 2024, at 10:55 AM, Steve 'dillo Okay <espre...@gmail.com> wrote:

The LuwuDynamics XGO-series of quadruped robots have hard nylon legs that are good for a product where you don't know what sort of surface somebody might use it on. In practice, they kind of suck for traction indoors. They're slippery and the robot has a shuffling,sliding kind of gait vs. a walk.

I tried gasket foam and trayliner mat, but those all either wore easily or didn't grip much(if any). So I did some CAD in OnShape, fired up the printer w/ a spool of TPU and a few hours later, little rubber dog-shoes!

It looks like this:

<PXL_20240708_041556647.jpg>

and works like this:
https://youtu.be/1I5Dhr9_iXA

--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/809123b4-57a4-4998-a5f8-37bda24d641dn%40googlegroups.com.
<PXL_20240708_041556647.jpg>

[Stephen Williams]

Friction tape used to have grit in the tape, so it would scratch floors.  Perhaps they stopped that.

One of the best elastomers for contact with the floor / ground or other friction purposes is urethane, like skate wheels use now.  All skate wheels I think as nothing else is competitive.  Grippy, available in a wide range of durometers, non-marking because it is very long wearing.  You can cast it with a mold of whatever you need.  Or you could find a way to use inline skate wheels or half wheels.

sdw

To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/0C2BCF65-EBC2-4D32-A031-23B16988E4B2%40gmail.com.

--

Stephen D. Williams Founder: VolksDroid, Blue Scholar Foundation 650-450-8649 | fax:703-995-0407 | s...@lg.net | https://VolksDroid.org | https://BlueScholar.org | https://sdw.st/in

[James H Phelan]

Dillo,

On my Freenove robot dog I used paintball rubber balls, drilled 1/4" hole 1/2 way through & stuck on my dog's feet:

https://www.amazon.com/dp/B08J62CBW7?ref=ppx_yo2ov_dt_b_product_details&th=1&psc=1

RoverDoc

James H Phelan
"Nihil est sine ratione cur potius sit quam non sit"
Leibniz

[Steve " 'dillo" Okay]

On Friday, July 12, 2024 at 3:47:28 PM UTC-6 James H Phelan wrote:

Dillo,

On my Freenove robot dog I used paintball rubber balls, drilled 1/4" hole 1/2 way through & stuck on my dog's feet:

https://www.amazon.com/dp/B08J62CBW7?ref=ppx_yo2ov_dt_b_product_details&th=1&psc=1

I looked up the Freenove and it certainly looks like it could use a better contact surface than those "peg-legs".

Are you using them as a round ball or did you cut them in half ?

I've noticed the ball shape is sort of the dominant "foot" shape for quadrupeds. 

I'm guessing that's because it seems to offer the optimal contact area on arbitrary surfaces, esp. when going over obstacles or uneven terrain.

Thanks for the info.

'dillo

[Chris Albertson]

With 3-DOF legs we can control thre location of the foot but not ther orientation of the foot releative to the ground.  A sphere is unique in that it is the only shape where orintation does not matter.    If ypoui want to use flatt feet, then you need an ankle.

Even on a flat floor, the ball shape has an advantage.  When the foot is plated on the floor and the robot moves, a round foot will roll, Any other shape woud rock the heel upward, unless there was an ankle.

--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/5602d07c-8d7e-4e74-b0ea-66dbd1a11ed0n%40googlegroups.com.

[Stephen Williams]

A swivel foot could be interesting.  Free-swiveling, it provides a lot more contact with the floor.  This or something like it could be modified for active engagement.  Could potentially use 2 per foot too.

https://www.amazon.com/Luomorgo-M10-Adjustable-Furniture-Mechanical/dp/B0BMV2K8V3

sdw

[Steve " 'dillo" Okay]

On Sunday, July 14, 2024 at 12:37:03 PM UTC-6 Chris Albertson wrote:

With 3-DOF legs we can control thre location of the foot but not ther orientation of the foot releative to the ground.  A sphere is unique in that it is the only shape where orintation does not matter.    If ypoui want to use flatt feet, then you need an ankle.

Even on a flat floor, the ball shape has an advantage.  When the foot is plated on the floor and the robot moves, a round foot will roll, Any other shape woud rock the heel upward, unless there was an ankle.

The ankle is considered a diagnostic feature of primates and many(including us) have a ball-shaped feature on the left-most toe which lets you pivot and roll off it, perfect for surmounting and rolling off uneven surfaces, or, say, clamboring amongst the trees. We(esp. those of us who sit in front of desks all day) don't use this that often, but it's there. You know, just on the off chance you might find yourself needing to climb on/off an autonomous vehicle to check on a LIDAR or 5G antenna or something :)

'dillo

 

[Chris Albertson]

On Jul 14, 2024, at 8:40 PM, 'Stephen Williams' via HomeBrew Robotics Club <hbrob...@googlegroups.com> wrote:

A swivel foot could be interesting.  Free-swiveling, it provides a lot more contact with the floor.  This or something like it could be modified for active engagement.  Could potentially use 2 per foot too.

First off a ball is the perfect shape.  It works well and there is no problem that needs to be fixed, other then the added mass of the ball.  A lighter solution would be better.  The problem is the added mass.  It not only makes the robot heavier but the moment of the added mass causes a reactive torque on the robot every time it moves it’s foot. and of course the moment at the end of the leg stress the motors joint moters and the legs moves slower then it otherwise would.   All this seem like a small detail but the end of the leg is the worst place to add mass to a quadruped.   Or said the other way, removing mass from the legs is the best why to impprove the perfromance of any quadruped. (this likey applies to bipeds too.)

That said, some features are worth the added mass. Number one on my list wouild be force sensors on the feet to measure the force on the floor on the bottoms of the feet.  Even basic on/off switches would help a lot.

If traction is an issue the first thing I’d try is to find a softer rubber ball, failing that try cutting (or printing) treads into the rubber, a tiny size waffel grid could work.  But traction will not be an issue if ther force on the floor is corrrect.  I’ve watched and vidio'd the feet and any time I see loss of traction it usually is because the robot walking in “open loop” mode and can’t sense the force of the foot on the ground so it is pushing backward with the foot still in the air.   It is really hard to compute where the floor is even if you get the math exactly right because the robot “flexes” and bends at the millimeter level and the kinimatic equations are always an approximation.   So closing the loop is worth the few grams it would take.

Note I say “balls” but they could be 3D printed hard platic covered in 3D printed TPU “tires”.  Either way you get a soft rubber hemispere.

Bipeds are different.  I have seen bipedal working using shperical feet, the robot must be very good at balance to do this.  But most bipods have articulated ankles at least in the fore/aft direction and pressure sensors in the feet.  Bipods also need to be able to apply torque the ground to rotate the body and a single-pint contact area would just spin in place.  This adds a lot of cost and weight but might be needed just for basic standing in place.

https://www.amazon.com/Luomorgo-M10-Adjustable-Furniture-Mechanical/dp/B0BMV2K8V3

sdw

On 7/14/24 11:36 AM, Chris Albertson wrote:

With 3-DOF legs we can control thre location of the foot but not ther orientation of the foot releative to the ground.  A sphere is unique in that it is the only shape where orintation does not matter.    If ypoui want to use flatt feet, then you need an ankle.

Even on a flat floor, the ball shape has an advantage.  When the foot is plated on the floor and the robot moves, a round foot will roll, Any other shape woud rock the heel upward, unless there was an ankle.

On Jul 14, 2024, at 9:38 AM, Steve 'dillo Okay <espre...@gmail.com> wrote:

On Friday, July 12, 2024 at 3:47:28 PM UTC-6 James H Phelan wrote:

Dillo,

On my Freenove robot dog I used paintball rubber balls, drilled 1/4" hole 1/2 way through & stuck on my dog's feet:

https://www.amazon.com/dp/B08J62CBW7?ref=ppx_yo2ov_dt_b_product_details&th=1&psc=1

I looked up the Freenove and it certainly looks like it could use a better contact surface than those "peg-legs".

Are you using them as a round ball or did you cut them in half ?

I've noticed the ball shape is sort of the dominant "foot" shape for quadrupeds. 

I'm guessing that's because it seems to offer the optimal contact area on arbitrary surfaces, esp. when going over obstacles or uneven terrain.

Thanks for the info.

'dillo

--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/0b59cc8e-8f6e-47c5-ba79-b1f615d96eba%40lig.net.