Wheel selection (1/2) - Front wheels

36 views
Skip to first unread message

David Crawley

unread,
Jan 28, 2013, 9:26:06 PM1/28/13
to hbrob...@googlegroups.com, maryff...@gmail.com, darka...@gmail.com, chr...@yahoo.com, jame...@hotmail.com, amit...@gmail.com, Kenny Klittich, busint...@gmail.com
I am convening a team out of the Hacker Dojo that is building a group designed Robot. We are building a differential drive floor robot, we expect the robot to have 2 front drive wheels and 2 rear casters. The intent is for the robot to be able to handle >>50lbs of payload, and be able to easily navigate anywhere that is ADA compliant (light outdoor use is implied). A drawing of the initial design of the robot is attached (design.jpg). We'd like the robot to be able to handle a 75mm (3") curb .

We have various options for our drive wheels. Please see attached. What are the trade offs on these wheels? What is everyone's vote as to which wheel we should use? So far I have been guided to wide wheels to allow for light off-road use (this came from a disabled wheelchair designer who spends his life making driving a differential drive system through the real world).

Looking forward to everyone's opinion.

David


design.jpg
Compare1&2.jpg
Front3a.jpg
Front3b.jpg
Frontdrive1a.jpg
Frontdrive1b.jpg
frontdrive2a.jpg
frontdrive2b.jpg

Butokim

unread,
Jan 28, 2013, 11:11:31 PM1/28/13
to hbrob...@googlegroups.com
David,
 
I have a Jet 2 electric wheel chair with 14" wheels and more power than I know what to do with and there is NO chance that it will pull up over a 3" curb...unless it is very very sloped.
 
Jim


From: hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com] On Behalf Of David Crawley
Sent: Monday, January 28, 2013 6:26 PM
To: hbrob...@googlegroups.com
Cc: maryff...@gmail.com; darka...@gmail.com; chr...@yahoo.com; jame...@hotmail.com; amit...@gmail.com; Kenny Klittich; busint...@gmail.com
Subject: [HBRobotics] Wheel selection (1/2) - Front wheels

anfederman@comcast

unread,
Jan 29, 2013, 11:03:13 AM1/29/13
to hbrob...@googlegroups.com
I concur Jim,  while a rule of thumb is the wheel diameter needs to be 3 to 4 times the highest vertical obstacle, in reality it needs to be about 5 times.  You would need a fifteen inch wheel to go over a 3 inch curb ( and also a lot of torque.) Think about it this way,  when rolling along the ground,  you start off overcoming inertia and friction, so you only need a fraction of the "F"  to move the "M" and get the "A."  When you climb a curb you are  lifting the full weight of the robot straight up and  that  is "M*G."  Fortunately,  there is a basic machine that solves this problem.
 
It is called the "inclined plane." Attaching a front wedge shaped skid plate is a simple solution and should solve the curb problem.  This approach involves getting the robot up to speed,  and jumping up onto the curb.  Probably not the best technique for pizza delivery.  More sophisticated solution exist, at he cost of complexity and extra wieght:
 
    1.  Active or passive suspension.
    2.  Wedge shaped tracks instead of wheels.
    3.  Active climbing assist system( Forklift or arm mechanism pushing on the ground.)
 
Please be advised that if you can climb up curbs, you must be prepared to climb down curbs in a manner that does not destroy the robot or spill the cargo all over the street. Winking smile emoticon  This is not a trivial problem.  How many home built stair climbing robots have you seen?
Emoticon3.gif

Patrick Goebel

unread,
Jan 29, 2013, 11:24:44 AM1/29/13
to hbrob...@googlegroups.com
In an earlier thread, Alan KM6VV wrote this about this 6WD bot climbing
stairs. So maybe 6 wheels are better than 4 for this application?

> Austin,
>
> What's the wheel diameter and ground clearance on the Slash?
>
> I'm working with 4907 T-Maxx suspension parts (6.3" tires). My 6WD
> custom chassis can climb stairs, and run through my front yard and
> on/off curbs and the driveway. I'm guessing (hoping) it's OK for
> Robomagellan.
>
> http://www.lynxmotion.net/viewtopic.php?f=9&t=7839&p=85121#p85121
>
>
> Alan KM6VV

Steven Nelson

unread,
Jan 29, 2013, 12:03:36 PM1/29/13
to hbrob...@googlegroups.com
A nice read with pictures... I like pictures..;)  
http://c3p0.ou.edu/IRL/Theses/Roman-MS.pdf


--
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+unsubscribe@googlegroups.com.
To post to this group, send email to hbrob...@googlegroups.com.
Visit this group at http://groups.google.com/group/hbrobotics?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.





--
Inspection time: The period of time required to inspect a recently welded piece of metal using your bare hand.
The hotter the metal the shorter the period of inspection time

KM6VV

unread,
Jan 29, 2013, 12:15:46 PM1/29/13
to hbrob...@googlegroups.com
I can climb curbs and steps with my 6WD RoboMagellan 'bot! 5.75" tires.
Independent suspension on each tire. I have a "skid plate" on the
front and rear, a 6" wide by 2" 25 degree angled 1/32" aluminum plate.

Here's a link to DPA's bot, upon which mine is modeled:

http://www.geology.smu.edu/~dpa-www/robo/jbot/jbot_steps_004.mpg

Alan KM6VV

On 1/29/2013 8:03 AM, anfederman@comcast wrote:
> I concur Jim, while a rule of thumb is the wheel diameter needs to be 3
> to 4 times the highest vertical obstacle, in reality it needs to be
> about 5 times. You would need a fifteen inch wheel to go over a 3 inch
> curb ( and also a lot of torque.) Think about it this way, when rolling
> along the ground, you start off overcoming inertia and friction, so you
> only need a fraction of the "F" to move the "M" and get the "A." When
> you climb a curb you are lifting the full weight of the robot straight
> up and that is "M*G." Fortunately, there is a basic machine that
> solves this problem.
> It is called the "inclined plane." Attaching a front wedge shaped skid
> plate is a simple solution and should solve the curb problem. This
> approach involves getting the robot up to speed, and jumping up onto
> the curb. Probably not the best technique for pizza delivery. More
> sophisticated solution exist, at he cost of complexity and extra wieght:
> 1. Active or passive suspension.
> 2. Wedge shaped tracks instead of wheels.
> 3. Active climbing assist system( Forklift or arm mechanism
> pushing on the ground.)
> Please be advised that if you can climb up curbs, you must be prepared
> to climb down curbs in a manner that does not destroy the robot or spill
> the cargo all over the street. Winking smile emoticon This is not a

Wayne C. Gramlich

unread,
Jan 29, 2013, 1:43:27 PM1/29/13
to hbrob...@googlegroups.com, Wayne C. Gramlich
All:

There is another pretty cool wheel arrangement called the
Tri-Star wheel:

http://en.wikipedia.org/wiki/Tri-star_(wheel_arrangement)

There is a book called "Android Design: Practical Approaches
for Robot Builders" by Martin Weinstein (1981) that dedicates
some a number of pages to the topic. My copy of Android Design
got loaned out and never found its way back to me. Used copies
are available from Amazon.

Regards,

-Wayne

On 01/29/2013 09:03 AM, Steven Nelson wrote:
> A nice read with pictures... I like pictures..;)
> http://c3p0.ou.edu/IRL/Theses/Roman-MS.pdf
>
> On Tue, Jan 29, 2013 at 8:24 AM, Patrick Goebel <pat...@pirobot.org
> <mailto:pat...@pirobot.org>> wrote:
>
> In an earlier thread, Alan KM6VV wrote this about this 6WD bot
> climbing stairs. So maybe 6 wheels are better than 4 for this
> application?
>
> Austin,
>
> What's the wheel diameter and ground clearance on the Slash?
>
> I'm working with 4907 T-Maxx suspension parts (6.3" tires). My
> 6WD custom chassis can climb stairs, and run through my front
> yard and on/off curbs and the driveway. I'm guessing (hoping)
> it's OK for Robomagellan.
>
> http://www.lynxmotion.net/__viewtopic.php?f=9&t=7839&p=__85121#p85121
> <http://www.lynxmotion.net/viewtopic.php?f=9&t=7839&p=85121#p85121>
>
>
> Alan KM6VV
>
>
> --
> 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+unsubscribe@__googlegroups.com
> <mailto:hbrobotics%2Bunsu...@googlegroups.com>.
> To post to this group, send email to hbrob...@googlegroups.com
> <mailto:hbrob...@googlegroups.com>.
> Visit this group at
> http://groups.google.com/__group/hbrobotics?hl=en
> <http://groups.google.com/group/hbrobotics?hl=en>.
> For more options, visit https://groups.google.com/__groups/opt_out
> <https://groups.google.com/groups/opt_out>.
>
>
>
>
>
> --
> Inspection time: The period of time required to inspect a recently
> welded piece of metal using your bare hand.
> The hotter the metal the shorter the period of inspection time
>
> --
> 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.

Wayne C. Gramlich

unread,
Jan 29, 2013, 2:00:04 PM1/29/13
to hbrob...@googlegroups.com, Wayne C. Gramlich
On 01/28/2013 06:26 PM, David Crawley wrote:
> I am convening a team out of the Hacker Dojo that is building a group
> designed Robot. We are building a differential drive floor robot, we
> expect the robot to have 2 front drive wheels and 2 rear casters. The
> intent is for the robot to be able to handle >>50lbs of payload, and be
> able to easily navigate anywhere that is ADA compliant (light outdoor
> use is implied). A drawing of the initial design of the robot is
> attached (design.jpg). We'd like the robot to be able to handle a 75mm
> (3") curb .

Other people are commenting on the curb issue.

The bigger problem you have is that by putting the two drive wheels
at the front, when your software decides to turn in place, the robot
will sweep out a fairly large area. Murphy's law as applied to robots
goes something like "A turning robot will knock over everything near
it." I know this from personal experience, since my large robot
platform is currently in the wheels forward format. The next iteration
will be pulling the wheels in somewhat.

The alternative is to pull the drive wheels closer to the center
of the robot, like the Roomba or Neato XV-11. This brings up
other issues about where to put the casters wheels, how many
caster wheels, how to prevent floor obstacles from lifting a
drive wheel, etc.

There is no perfect wheel arrangement. There is a range of
trade-offs

> We have various options for our drive wheels. Please see attached. What
> are the trade offs on these wheels? What is everyone's vote as to which
> wheel we should use? So far I have been guided to wide wheels to allow
> for light off-road use (this came from a disabled wheelchair designer
> who spends his life making driving a differential drive system through
> the real world).

I would use the wheel that does not leave skid marks all over the
floor. Black rubber is out. After that go for the wheel that
gives the best floor traction. Wheel slippage causes grief.

Regards,

-Wayne

Butokim

unread,
Jan 29, 2013, 2:22:23 PM1/29/13
to hbrob...@googlegroups.com
I'm sure I am showing my ignorance but how does one do any kind of odometry
stuff with 6 wheels and all the (intentional) slipping?

James M. Geidl, K6JMG
D.B. Cooper, you have a message.



-----Original Message-----
From: hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com] On

KM6VV

unread,
Jan 29, 2013, 2:49:41 PM1/29/13
to hbrob...@googlegroups.com
Yeah, odometry can be rough.

Solution is to make odometry a part of a larger navigation solution.
Accelerometers, compass, GPS and dead-reckoning all have a contribution
to make. "Sensor Fusion" can unite these bits of data, and give a
better estimation of position, heading and velocity.

Alan KM6VV

anfederman@comcast

unread,
Jan 29, 2013, 7:38:24 PM1/29/13
to hbrob...@googlegroups.com
You can't use wheel odometry reliably on a skid steer. Better make that you
can't use it reliably period. Even with two wheels, there will be slippage
and inaccuracy.

You can use mms ( accelerometers, rate turning, compass and other inertial
guidance) and combine that with other distance sensors. Low end solutions
use SLAM (based on something like the Kinnext or XV-11 LDS) combined with
wheel odometry, but I am not sure how much is based on what the wheels do.

--------------------------------------------------
From: "Butokim" <but...@gmail.com>
Sent: Tuesday, January 29, 2013 11:22 AM
To: <hbrob...@googlegroups.com>
Subject: RE: [HBRobotics] Wheel selection (1/2) - Front wheels
Reply all
Reply to author
Forward
0 new messages