I'll see what I can generate as far as pictures (though I've never attached anything to a post before, but it can't be too hard). FYI I love documenting in POWER POINT (PPT not PPTX). Anyone that could read it and repost as PDF would be helpful.
As far as things "on top of a rubber ball" I, more than I would like to say, have design issues. A 9 DOF would b great but like others there are FUNDING issues. The parts boxes had ball bearing, wires and copper tape. The broken sweeper had a nice plastic ring (tough this could have been a piece of ABS plastic). The ball bearing is attached to the wire (solder & glue). The wire hangs from a modified coat hanger, that towers above the plastic ring. The longer the wire the smaller the angle of movement. This technique works for static balance an at best a slow-creep movement across the floor. The ball-wire is ground. The eight copper tape segment wraps have 10K pull-ups. Each of the pulled up segments goes to a BS2 input (propeller a little bit later in the project). Ball touches, signal goes low. Ball taps (small duration) like in just minor bouncing, might want to do a little sower movement and prepare some counter movement. Ball stays (longer duration) like in just about to fall over, need to stop movement and do some real motor counter balancing movement. NEED TO KEEP THE TRAINING WHEELS ON UNTIL YOU HAVE A GOOD HANDLE ON THE SOFTWARE. (my training wheels are just coat hangers bent out like fingers)
The wow-wee works great on the floor without the single-point-of-contact-ball when trying to develop some code to handle simple ball-wire signal detection and how much movement the slow/static system can handle.
Now is this what I plan on continuing with? NO NO NO NO NO. Santa promised I could get a real 9 DOF next month (a bit early for Christmas) that WILL WORK WITH A PROPELLER (3v not 5v).
......................+
+
wire .............. |............... | coat hanger
.......................|............... |
.......................|............... |
ball ............ +++............. |
bearing ......+++............. .|
...................+++.............. |
RING .......*********.......... |
.................*********.......... |
.................*********.......... |
============================= SNAP plate on top of
WOW-WEE motor assembly
(WOW-WEE head removed)
want smaller tilt angle: increase tower height, or move the lego's in
want an ALMOST TILTED and a "REACT NOW TILT", make two lego towers, one with legos in towards the center (ALMOST) and one with legos farther out (REACT NOW).
Components: LEGOs, copper tape (Sparkfun has it), stranded wire, and a center weight
Note I did drill a hole in the top plate to let the center wire down.
This circuit is only good for static balance development not dynamic balance while moving.
The circuit contacts have up to 2K resistance (after a couple days of air exposure), so the pull-up needs to be 10K.
A TAP means starting to lean, A CONSTANT CONTACT mean tilting.
My guess is that you (Duane) already have a 9DOF (or better), and I've seen your tri-motor base. So the question is "have you static or dynamic balanced it on a ball?". Static to me is just ball balance without trying to move somewhere. Dynamic is trying to move on a course or even trying to overcome obstacle's (like pencils or lego blocks).
My guess is that you (Duane) already have a 9DOF (or better), and I've seen your tri-motor base. So the question is "have you static or dynamic balanced it on a ball?".
As if I'd be able to keep such a feat a secret. Whenever I do something which might be considered cool (related to robots or microcontrollers) I'm quick to let everyone know about it.
I have thought about attempting something like this many times but I have yet to put much effort in attempting to balance a bot on a ball. While three omni wheels seems like the most logical wheel configuration, I've also wondered about Mecanum wheeled robots and even hexapods as candidates for ball balancing bots.
Duane, I'm not trying to challenge anyone but myself. I do a lot of research and try not to make the same mistakes others have already published. Three wheels should work. One wheel on a ball is more of a challenge than I'm ready for currently (nothing is impossible).
Some interesting results from last nights testing:
Cut the bottom off a Styrofoam cup, added three strips (sensors) of copper tape to the inside of the remaining cup, spaced out so they would each be ABOVE a wheel. Put a copper circle on the bottom inside of the cup. Added wires to all four contacts. Made the center circle ground. Added a ball bearing from a skate-board to roll around the cup. Each of the sensors had a 10K pull-up. When one sensor went to ground (ball touching it because it was tilted low) I drove the OTHER TWO wheels to regain balance. NOTE the ball could not touch two sensors at any point in time (spacing), so there were a few small angles that could result in a FOBR.
Duane, thank you for the DOF (IMU) suggestions. I want to do a bit more research before I purchase as I've heard some units have NOISE issues.
Couple of Questions:
1. Does anyone have a datasheet for "350-00029 850NM T1 3/4 Phototransistor, This phototransistor responds to visible light as well as the lower infrared range" ?
2. I was wondering if I had a little heavy steel ball inside my air filled big ball, would it help the balance standing still issue?
I want to do a bit more research before I purchase as I've heard some units have NOISE issues.
I haven't used the 10 DOF board but I have used the other two boards I mention. The HMC5883L I used with the Prop and I was amazed how steady a ready I could get (I did have trouble when using it near servos). I used the 6 DOF sensor with an Arduino and was convinced black magic was at work in order for something so inexpensive to work so well. The $3 6 DOF board was eerily stable.
I don't think the ball bearing tricks will work well. It won't be able to distinguish the force from acceleration from the force from gravity. You really need a gyro too.
I've replaced the styro-cup with a 2" PVC CAP (I will refer to as the bearing-cup). I drilled a hole in the bottom of the cap for a mounting bolt and ground connection. I lined the bottom inside of he cap with copper tape. I added SIX (Three wide and Three narrow) strips of tape to the inside of the circular side. The WIDE strips (sensors) are inline with the tri-bot wheel locations. The narrow strips (sensors) are in the in-between-wheel locations. The curved bottom of the cap helps pull the ball bearing back to center when the system is in BALANCE. It looks good enough to remove the training wheels off the tri-bot.
The bearing-cup unit will be used to help with the 9-DOF development (just as the training wheels were helpful with just basic balance). I plan on allowing the bearing-cup to be the OVER-RIDE for balance (too much noise on the 9-DF, situation the 9-DOF can't handle, or jut dumb "why didn't I think of that" software design).
Two areas that comments / suggestions are welcome:
1. What is the max grade (slanted floor) that EVE (IT) should test to?
2. What is the max obstacle (example: entrance way door threshold) that EVE (IT) should test to?
I have not seen a single ball bot that was tested to item #2.
It would be nice if the center of gravity (CG) was a bit lower. I've noticed several Arduino projects that put the motors and stuff into a rolling hamster ball. What if the MOVEMENT (M) motors and batteries were in the ball and you could command them via di-directional IR (BDIR), oh and the BALANCE and DESIRE (B-D) stuff was on top of the ball (current tri-bot concept)?
B-D would decide to move and command M to go.
As M moves B-D would adjust to stay on top (gyro, compass, acceler...)
As M is moving, B-D could command a LEFT, RIGHT, FORWARD, REVERSE, FAST, SLOW, STOP
The BDIR would provide: confirmation of command, AND allow B-D to know what orientation M was in.
The BDIR would be at least three sets of IRs. So when M sent a message (response, whatever) up to B-D using TX-1 then TX-2 and finally TX-3, B-D could associate a current M TX-n to it's RX-n and thus know which orientation M was in.
MOTOR circuit noise and weight would be in the lower CG ball.
Wouldn't this double the complexity? Doesn't the B-D stuff have to have its own motors to stay balanced? So you'd have motors in the ball and on top of the ball?
YES TWO separate motor systems. The bottom one should have heavy powerful motors / batteries as this is where the CG is. The top (B-D) only needs to worry about balance an may have light weight electronics.
This might help the software design as there is no augmenting motor circuits with balance demands and movement demands.
It's a little spider running on top of a big bowling ball (assuming the bowling ball knows where it I going, different then my bowling experiences). The spider tells the bowling ball what to do and then does whatever it (t spider) needs to do just to stay on top (otherwise it may be run over).
One of the test issues was the need for friction so the tri-bot could FORCE the ball to move (see bowling ball FOBR). Another concern was "IF I GET IT T MOVE, HOW DO IT STOP IT". Even if I have enough friction I might have a problem stopping a freight train.
Now (using two motor systems):
1. friction, CG, total mass are concerns for the lower ball system
2. direction and balance are concerns for the upper system
The Bear never stops moving after stepping onto the ball. Each leg adding weight to the side of the ball that needs it at that moment.
A 'Grippy' rubber ball, "in theory", could/should be able to climb an obstacle that is Less than half the height of the rubber ball.
A smooth Billiard ball might not have the traction needed.
Interesting to me, is that the ball stops when the bear performs the turn around.
The turn around was done by the Bear, and not the ball.
It's an interesting study, I learned that Bears are heavy, and agile too...
I think each of the Bears feet have a built in 'Incline' detector, and each foot knows when it is going uphill or downhill.
While the Bears head remains relatively level to the floor, all the paws are very busy going up and down hill...errr,. ball...
the obstacle issue is a problem, at least for me and my test (not finalized design) equipment. When the unit (EVE or IT, whatever you want to call the thing) comes up to a door threshold forward momentum stops and the balance stuff goes into crisis oscillations just fore a second or two, then it is still balanced on top but just sitting there (looking dumb at me while I look equally dumb at it).
I agree that the rubber ball should have enough GRIP to climb over the threshold but when I try to force drive the motor circuits (ignoring the balance demands) it just throws the tri-bot off. I think the BEAR has more mass then my tri-bot (and a 9-dof somewhere).
Unless the two motor system (B-D + M) can overcome this issue, then I think I will be forced to get one or more 9-DOFs and a bunch more mass for the single (above the ball) system. The single system will require complex action-reaction to the movement (slowing down while going up the threshold and speeding up while going down the threshold).
Do you think we can ask the BEAR to show a example of moving the ball across a small log (the bear has a big ball so the obstacle should be scaled up)?
What if we added an extendable counter-weight, that could be extended over/near the threshold when trying to move onto it, and maintain that location when trying to move off the threshold?
For sure, putting the weight where it needs to be is going to be the biggest challenge.
When climbing an obstacle with a ball, you have to lift the weight of the ball and the passenger.
Scaling the obstacle to the size of the ball will also be necessary of course..,
What if we added an extendable counter-weight, that could be extended over/near the threshold when trying to move onto it, and maintain that location when trying to move off the threshold?
No need for that, the top of the bot would just lean into a bump until it clears it, then lean backwards briefly. An obstacle/bump is just like a small hill. All Segways and balance bots have to lean into hill to climb. Watch this little bot lean to go up & down hills at 2:40 and 3:30:
WOW I would have never though of that, using a propeller augmented parakeet. Lets see (using posting #1 requirements, modification of 1 13 and 14, eliminating 4):
1.--She COULD fly as long as it's not dangerous
2.She must use at least one Parallax microcontroller (this is a Parallax forum). ******
3.She must have only one point of contact with the ground and balance while moving.
4.
5.She can be as tall or short as practical (some misproportions OK).
6.She must have one or more LCDs or other LED arrangement for emoting. ******************
7.She can have a wireless communications port.
8.She must have a storage door and storage compartment.
9.She must be able to detect some target object.
10.She must have arms that move.
11.She can have openings for object detection sensors.
12.Her exterior must be wireless except for attaching body parts.
13.--She must use batteries and can be rechargable. OR ALTERNATE FUEL SOURCE (seeds, ...)
14.--She can be programmed in any lanugage. (POLLY WANT A CRACKER....)
15.She can be made of any material.
16.She should be able to move her head in some way.
so -Tommy all you have remaining are requirements 2 and 6. Any idea of where you will mount the propeller and LED?
Erco I think the two wheeled balancing bot did have some reaction time issues (see when it is on the glass plate that is tilting back and forth under it).
Even though it was able to go up a hill and back down with "lean", I wonder if it's reaction time would allow movement across the door threshold.
My threshold was purchased at Home Depot and has a quarter circle entrance and exit ramp. It is not an inclined plane as seen in the bot video.
Should the testing for obstacles be an inclined plane, if so what is the maximum degree?
Where do you find a two piece clear plastic ball? The pet store hamster balls are too small for the tri-bot base.
My first attempt is using two plastic salad bowls. The initial results looked promising as the tri-bot could stay balanced while the Salad-Bowl Base (SBB) was moving. The tri-bot could also rotate the upper body (while maintaining balance) independent of the SBB.
The bad thing about SBB is
1. It was smoother on the outside then I wanted
2. I had to glue the pieces together (and thus break them after testing)
3. The IR bi-directional communications was sporadic
Any suggestion for a two piece BIG clear plastic ball? (maybe I'll get more response then I get on my other postings)
Where do you find a two piece clear plastic ball? The pet store hamster balls are too small for the tri-bot base.
My first attempt is using two plastic salad bowls. The initial results looked promising as the tri-bot could stay balanced while the Salad-Bowl Base (SBB) was moving. The tri-bot could also rotate the upper body (while maintaining balance) independent of the SBB.
The bad thing about SBB is
1. It was smoother on the outside then I wanted
2. I had to glue the pieces together (and thus break them after testing)
3. The IR bi-directional communications was sporadic
Any suggestion for a two piece BIG clear plastic ball? (maybe I'll get more response then I get on my other postings)
I want to know where the photos and video of this thing are? You're keeping all the fun to yourself!
If I was keeping the fun to myself, I WOULD NOT POST any findings.
I have two hands and only one tri-bot base (with a bunch of wires connected to the tri-bot motor control board).
If I video I can not save the tri-bot when the ball gets mad and tries to throw it off (or as in several tests, the wire bundle gets hung up).
In several postings I have asked for response and got absolutely NOTHING (a datasheet, a max incline angle, plastic ball sources).
In one posting I included a picture of a LEGO version of a simple 2D tilt detection sensor, trying to share an idea. I did not post the picture of the eight contact modified sweeper part that no-one-else would have tried to make.
Pictures / video are a form of sharing but it needs to be a two-way street.
Hang in there GrandeNurse, Some folks can't post everyday, and it takes time for some threads to gain momentum, this thread is already 3 pages long.
This is an interesting thread because of your contributions. Thank you for your time.
Don't worry, There will be more folks checking in soon enough. You already have Duane's attention, and that is usually a good thing.
I wish I could build a copy of your 8 sensor thingy. sounds like it could be used for more than just ball walking.
But for now, I am in the middle of building my dream workshop, it's going to take a while, because it happens to be getting built on top of my last dream workshop...
I don't get to play with Robots of any kind right now.
Anyway, you are doing just fine, If you don't want to post video, that's okay, we are all still interested in what you are doing, keep it up!.
Tommy, if you are building a workshop I would suggest adding a strong center beam. I added one to my shop with a lift rail & roller, it now serves the following:
1. When you are ready to paint (stain, polyU, finish pain coat, ...) it provides a place to HANG pieces (and newspaper to cover the floor from drips).
2. When you are working on repairing the SAMSUNG plasma wall TV (again, bad design with marginal caps), it allows you to hang it from the rail so you can get to both sides.
3. When you need to change out the rear tire of your Honda Goldwing, it allows you to wench it up without trying to use side jacks (that always get in the way).
4. When you are building tall boy dressers, go get a perfect plumb-bob line to see how true it is (assuming your ground just the center section of concrete LEVEL). You also get a strong helping arm when your truing up the frame.
5. Instead of using a welding table, you can suspend a working frame to clamp onto (when using welders, torches, and cutters), and slide it around to where you need it.
6. Assuming you use a strong I-beam from side to side of the workshop, it also makes a good place to mount power conduit & outlets, and air compressor outlets (though I've never got around the air part).
You will find the beam to be the helper you need when others are just not available (see the book Working Alone).
Now back to the subject: Does EVE (IT) have any requirement to move over terrain that might have bumps (like door thresholds), gullies, or mounds (tiny hill)?
Comments
As far as things "on top of a rubber ball" I, more than I would like to say, have design issues. A 9 DOF would b great but like others there are FUNDING issues. The parts boxes had ball bearing, wires and copper tape. The broken sweeper had a nice plastic ring (tough this could have been a piece of ABS plastic). The ball bearing is attached to the wire (solder & glue). The wire hangs from a modified coat hanger, that towers above the plastic ring. The longer the wire the smaller the angle of movement. This technique works for static balance an at best a slow-creep movement across the floor. The ball-wire is ground. The eight copper tape segment wraps have 10K pull-ups. Each of the pulled up segments goes to a BS2 input (propeller a little bit later in the project). Ball touches, signal goes low. Ball taps (small duration) like in just minor bouncing, might want to do a little sower movement and prepare some counter movement. Ball stays (longer duration) like in just about to fall over, need to stop movement and do some real motor counter balancing movement. NEED TO KEEP THE TRAINING WHEELS ON UNTIL YOU HAVE A GOOD HANDLE ON THE SOFTWARE. (my training wheels are just coat hangers bent out like fingers)
The wow-wee works great on the floor without the single-point-of-contact-ball when trying to develop some code to handle simple ball-wire signal detection and how much movement the slow/static system can handle.
Now is this what I plan on continuing with? NO NO NO NO NO. Santa promised I could get a real 9 DOF next month (a bit early for Christmas) that WILL WORK WITH A PROPELLER (3v not 5v).
......................+
+
wire .............. |............... | coat hanger
.......................|............... |
.......................|............... |
ball ............ +++............. |
bearing ......+++............. .|
...................+++.............. |
RING .......*********.......... |
.................*********.......... |
.................*********.......... |
============================= SNAP plate on top of
WOW-WEE motor assembly
(WOW-WEE head removed)
want smaller tilt angle: increase tower height, or move the lego's in
want an ALMOST TILTED and a "REACT NOW TILT", make two lego towers, one with legos in towards the center (ALMOST) and one with legos farther out (REACT NOW).
Components: LEGOs, copper tape (Sparkfun has it), stranded wire, and a center weight
Note I did drill a hole in the top plate to let the center wire down.
This circuit is only good for static balance development not dynamic balance while moving.
The circuit contacts have up to 2K resistance (after a couple days of air exposure), so the pull-up needs to be 10K.
A TAP means starting to lean, A CONSTANT CONTACT mean tilting.
Hope this helps until you can afford a 9 DOF
You can purchase a 6dof board for less the $3 shipped and a three axis magnetometer costs less. Of you can splurge and get a 10dof board for $8.35 shipped. All these sensors work just fine with the Propeller. I haven't found a 5V sensor which won't communicate just fine with the Prop.
I attempted to explain a bit about omni wheel math on this page.
There's some Spin code for controlling the motor speeds at the bottom of my omni wheel robot page. Make sure and watch the "SpinningStraight" video.
As if I'd be able to keep such a feat a secret. Whenever I do something which might be considered cool (related to robots or microcontrollers) I'm quick to let everyone know about it.
I have thought about attempting something like this many times but I have yet to put much effort in attempting to balance a bot on a ball. While three omni wheels seems like the most logical wheel configuration, I've also wondered about Mecanum wheeled robots and even hexapods as candidates for ball balancing bots.
Balancing on a ball has been done. Think big. You need to balance on a ball on top of another ball. Now you're barking! Man up!
Ten bucks says you can't do it before midnite!
Some interesting results from last nights testing:
Cut the bottom off a Styrofoam cup, added three strips (sensors) of copper tape to the inside of the remaining cup, spaced out so they would each be ABOVE a wheel. Put a copper circle on the bottom inside of the cup. Added wires to all four contacts. Made the center circle ground. Added a ball bearing from a skate-board to roll around the cup. Each of the sensors had a 10K pull-up. When one sensor went to ground (ball touching it because it was tilted low) I drove the OTHER TWO wheels to regain balance. NOTE the ball could not touch two sensors at any point in time (spacing), so there were a few small angles that could result in a FOBR.
Duane, thank you for the DOF (IMU) suggestions. I want to do a bit more research before I purchase as I've heard some units have NOISE issues.
Couple of Questions:
1. Does anyone have a datasheet for "350-00029 850NM T1 3/4 Phototransistor, This phototransistor responds to visible light as well as the lower infrared range" ?
2. I was wondering if I had a little heavy steel ball inside my air filled big ball, would it help the balance standing still issue?
I haven't used the 10 DOF board but I have used the other two boards I mention. The HMC5883L I used with the Prop and I was amazed how steady a ready I could get (I did have trouble when using it near servos). I used the 6 DOF sensor with an Arduino and was convinced black magic was at work in order for something so inexpensive to work so well. The $3 6 DOF board was eerily stable.
I don't think the ball bearing tricks will work well. It won't be able to distinguish the force from acceleration from the force from gravity. You really need a gyro too.
The bearing-cup unit will be used to help with the 9-DOF development (just as the training wheels were helpful with just basic balance). I plan on allowing the bearing-cup to be the OVER-RIDE for balance (too much noise on the 9-DF, situation the 9-DOF can't handle, or jut dumb "why didn't I think of that" software design).
Two areas that comments / suggestions are welcome:
1. What is the max grade (slanted floor) that EVE (IT) should test to?
2. What is the max obstacle (example: entrance way door threshold) that EVE (IT) should test to?
I have not seen a single ball bot that was tested to item #2.
B-D would decide to move and command M to go.
As M moves B-D would adjust to stay on top (gyro, compass, acceler...)
As M is moving, B-D could command a LEFT, RIGHT, FORWARD, REVERSE, FAST, SLOW, STOP
The BDIR would provide: confirmation of command, AND allow B-D to know what orientation M was in.
The BDIR would be at least three sets of IRs. So when M sent a message (response, whatever) up to B-D using TX-1 then TX-2 and finally TX-3, B-D could associate a current M TX-n to it's RX-n and thus know which orientation M was in.
MOTOR circuit noise and weight would be in the lower CG ball.
Any comments?
Wouldn't this double the complexity? Doesn't the B-D stuff have to have its own motors to stay balanced? So you'd have motors in the ball and on top of the ball?
This might help the software design as there is no augmenting motor circuits with balance demands and movement demands.
It's a little spider running on top of a big bowling ball (assuming the bowling ball knows where it I going, different then my bowling experiences). The spider tells the bowling ball what to do and then does whatever it (t spider) needs to do just to stay on top (otherwise it may be run over).
One of the test issues was the need for friction so the tri-bot could FORCE the ball to move (see bowling ball FOBR). Another concern was "IF I GET IT T MOVE, HOW DO IT STOP IT". Even if I have enough friction I might have a problem stopping a freight train.
Now (using two motor systems):
1. friction, CG, total mass are concerns for the lower ball system
2. direction and balance are concerns for the upper system
http://spectrum.ieee.org/automaton/robotics/military-robots/nasas-squishable-ball-bot-could-explore-titan
http://phys.org/news/2013-10-gimball-crash-happy-robot-video.html
http://www.neweggflash.com/product/N82E16886118001
http://www.toyrc.com/happy-cow-777-268-infrared-3ch-ir-ufo-flying-ball-w-gyro-black.html
http://spectrum.ieee.org/automaton/robotics/military-robots/japanese-ministry-of-defense-spends-1000-on-flying-robot-soccer-ball
http://www.simplebotics.com/2014/05/the-5-best-quadcopter-drones-for.html
http://article.wn.com/view/2013/02/25/Watch_This_Amazing_Robot_Cube_Deftly_Balance_Itself_Like_a_C/
the cubli: a cube that can jump up, balance and walk
and a HEXPOD / ball
http://www.popsci.com/technology/article/2011-12/video-graceful-globular-hexapod-bot-unfurls-leafy-limbs-and-scampers-along?dom=PSC&loc=recent&lnk=2&con=read-full-story
be really great if you could mix the "Amazing Robot Cube" and the "Hexpod". Didn't they make a STARWARS movie with something like that?
Can they get it across a entrance way door threshold obstacle ?
They implanted just the bear necessities.
The Bear never stops moving after stepping onto the ball. Each leg adding weight to the side of the ball that needs it at that moment.
A 'Grippy' rubber ball, "in theory", could/should be able to climb an obstacle that is Less than half the height of the rubber ball.
A smooth Billiard ball might not have the traction needed.
Interesting to me, is that the ball stops when the bear performs the turn around.
The turn around was done by the Bear, and not the ball.
It's an interesting study, I learned that Bears are heavy, and agile too...
I think each of the Bears feet have a built in 'Incline' detector, and each foot knows when it is going uphill or downhill.
While the Bears head remains relatively level to the floor, all the paws are very busy going up and down hill...errr,. ball...
-Tommy
I agree that the rubber ball should have enough GRIP to climb over the threshold but when I try to force drive the motor circuits (ignoring the balance demands) it just throws the tri-bot off. I think the BEAR has more mass then my tri-bot (and a 9-dof somewhere).
Unless the two motor system (B-D + M) can overcome this issue, then I think I will be forced to get one or more 9-DOFs and a bunch more mass for the single (above the ball) system. The single system will require complex action-reaction to the movement (slowing down while going up the threshold and speeding up while going down the threshold).
Do you think we can ask the BEAR to show a example of moving the ball across a small log (the bear has a big ball so the obstacle should be scaled up)?
What if we added an extendable counter-weight, that could be extended over/near the threshold when trying to move onto it, and maintain that location when trying to move off the threshold?
When climbing an obstacle with a ball, you have to lift the weight of the ball and the passenger.
Scaling the obstacle to the size of the ball will also be necessary of course..,
-Tommy
No need for that, the top of the bot would just lean into a bump until it clears it, then lean backwards briefly. An obstacle/bump is just like a small hill. All Segways and balance bots have to lean into hill to climb. Watch this little bot lean to go up & down hills at 2:40 and 3:30:
1.--She COULD fly as long as it's not dangerous
2.She must use at least one Parallax microcontroller (this is a Parallax forum). ******
3.She must have only one point of contact with the ground and balance while moving.
4.
5.She can be as tall or short as practical (some misproportions OK).
6.She must have one or more LCDs or other LED arrangement for emoting. ******************
7.She can have a wireless communications port.
8.She must have a storage door and storage compartment.
9.She must be able to detect some target object.
10.She must have arms that move.
11.She can have openings for object detection sensors.
12.Her exterior must be wireless except for attaching body parts.
13.--She must use batteries and can be rechargable. OR ALTERNATE FUEL SOURCE (seeds, ...)
14.--She can be programmed in any lanugage. (POLLY WANT A CRACKER....)
15.She can be made of any material.
16.She should be able to move her head in some way.
so -Tommy all you have remaining are requirements 2 and 6. Any idea of where you will mount the propeller and LED?
Even though it was able to go up a hill and back down with "lean", I wonder if it's reaction time would allow movement across the door threshold.
My threshold was purchased at Home Depot and has a quarter circle entrance and exit ramp. It is not an inclined plane as seen in the bot video.
Should the testing for obstacles be an inclined plane, if so what is the maximum degree?
My first attempt is using two plastic salad bowls. The initial results looked promising as the tri-bot could stay balanced while the Salad-Bowl Base (SBB) was moving. The tri-bot could also rotate the upper body (while maintaining balance) independent of the SBB.
The bad thing about SBB is
1. It was smoother on the outside then I wanted
2. I had to glue the pieces together (and thus break them after testing)
3. The IR bi-directional communications was sporadic
Any suggestion for a two piece BIG clear plastic ball? (maybe I'll get more response then I get on my other postings)
I want to know where the photos and video of this thing are? You're keeping all the fun to yourself!
I have two hands and only one tri-bot base (with a bunch of wires connected to the tri-bot motor control board).
If I video I can not save the tri-bot when the ball gets mad and tries to throw it off (or as in several tests, the wire bundle gets hung up).
In several postings I have asked for response and got absolutely NOTHING (a datasheet, a max incline angle, plastic ball sources).
In one posting I included a picture of a LEGO version of a simple 2D tilt detection sensor, trying to share an idea. I did not post the picture of the eight contact modified sweeper part that no-one-else would have tried to make.
Pictures / video are a form of sharing but it needs to be a two-way street.
This is an interesting thread because of your contributions. Thank you for your time.
Don't worry, There will be more folks checking in soon enough. You already have Duane's attention, and that is usually a good thing.
I wish I could build a copy of your 8 sensor thingy. sounds like it could be used for more than just ball walking.
But for now, I am in the middle of building my dream workshop, it's going to take a while, because it happens to be getting built on top of my last dream workshop...
I don't get to play with Robots of any kind right now.
Anyway, you are doing just fine, If you don't want to post video, that's okay, we are all still interested in what you are doing, keep it up!.
-Tommy
Don't you think Murata cheer leading robots satisfy most of EVE requirements ?
[video=youtube_share;KNi_WLYoRNY]
Kaeru no Ojisan
1. When you are ready to paint (stain, polyU, finish pain coat, ...) it provides a place to HANG pieces (and newspaper to cover the floor from drips).
2. When you are working on repairing the SAMSUNG plasma wall TV (again, bad design with marginal caps), it allows you to hang it from the rail so you can get to both sides.
3. When you need to change out the rear tire of your Honda Goldwing, it allows you to wench it up without trying to use side jacks (that always get in the way).
4. When you are building tall boy dressers, go get a perfect plumb-bob line to see how true it is (assuming your ground just the center section of concrete LEVEL). You also get a strong helping arm when your truing up the frame.
5. Instead of using a welding table, you can suspend a working frame to clamp onto (when using welders, torches, and cutters), and slide it around to where you need it.
6. Assuming you use a strong I-beam from side to side of the workshop, it also makes a good place to mount power conduit & outlets, and air compressor outlets (though I've never got around the air part).
You will find the beam to be the helper you need when others are just not available (see the book Working Alone).
Now back to the subject: Does EVE (IT) have any requirement to move over terrain that might have bumps (like door thresholds), gullies, or mounds (tiny hill)?
Mostly. But it wasn't done by a forum contributor ... yet.
Even entertainment is useful if it's not talking about celebs, reality, and who's cheatin' who.