RoboSavvy Forum

A day in the life of a robot

2006-07-31T13:25:48+01:00

well, the "GETMOTORSET" command reads the servos current position and keeps the servos in that position. This comand is used to make sure the robot is damaged, say for instance all the zero positions of your robot are 100, but if they all snap to that 100 really quick when turned on and your robot was in an odd position then a leg might rub an arm etc... and a wire could end up getting cut.

The "1"s and "0"s refer to whether or not the servos stay in position. In the case of a 1 the servo is read and stays in position, and in the case of a 0 the servo is read and set to the default position.

I hope that helps, and I am glad to hear that your like the routine i made beermat, you are right, i am pleased to hear that my routine works well with your robot, thats great! i personally enjoy the slight extra friction of those mats and the padding is always a plus with moves like these (espeially in the development stages).

Statistics: Posted by hivemind — Mon Jul 31, 2006 1:25 pm

A day in the life of a robot

2006-07-31T10:20:59+01:00

Hi Hivemind

What a perfect jump - I'm new to RN and eagerly watching for some cool moves - and you just saved the day!
I stored the move in K25 so now I'm jumping around the living room doing perfect jumps. On my entre carpet, he's always falling so I guess that the carpet (material/thickness etc) as you mention does have a great influence on the performance.

p.s. when reading the manual and looking at this line i'm confused of what the ones and zeroes does - can you help?
GETMOTORSET G24,1,1,1,1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,1,1,1,1,1,0

-Lars (Viking)

Statistics: Posted by Viking — Mon Jul 31, 2006 10:20 am

A day in the life of a robot

2006-07-31T02:44:26+01:00

lol

Statistics: Posted by beermat — Mon Jul 31, 2006 2:44 am

A day in the life of a robot

2006-07-31T02:25:11+01:00

Thats funny we use those mats to insulate beer kegs so they stay cool when sitting on the floor, beer mats.

Statistics: Posted by Bullit — Mon Jul 31, 2006 2:25 am

A day in the life of a robot

2006-07-31T02:02:35+01:00

Hive - I was in Sears today, and saw what appear to be the same interlocking foam mats that you have, so I grabbed a set. You'll be pleased to know that, apart from ending the handspring facing off-center, my RN does your (unmodified) handspring really well on the mat.

The version I was working on - and testing on hardwood floors - is now horrible on the mat, so it seems to make a suprising difference for this move.

Statistics: Posted by beermat — Mon Jul 31, 2006 2:02 am

A day in the life of a robot

2006-07-30T15:16:26+01:00

Exactly, that is just a loop, the cool part is that it remains stable, in fact all i had to do was bring the camera down and video, no retakes, just pure stable robot goodness.

I just wanted to show that the flip was stable and could be done multiple times without falling over because anyone who has tried it may not have gotten to work 100% yet, but dont be afraid of using and tweaking my code, the most important parts are honestly the delays. just make sure you have a decent zero position for your robot and it can do what Thor can.

robodude666, if you could create a small pro bmx bike for my robot i will make you some code for it to ride it

Statistics: Posted by hivemind — Sun Jul 30, 2006 3:16 pm

A day in the life of a robot

2006-07-30T15:03:44+01:00

Isn't that just doing a for loop to run the code 3 times? Now, try making it do the front handspring with the hip mods, and then go for a back handspring! Think you can make it ride pro in bmx after that?

Statistics: Posted by robodude666 — Sun Jul 30, 2006 3:03 pm

A day in the life of a robot

2006-07-30T14:49:45+01:00

After years of training (hours.. i just hadny put the video together) and lots of hard work (pushups) Thor is now able to do many flips in a row, on and off-terrain, can get back up and still has a smile on his face (the green light indicating that the accelerometer is good).

So here you are, youll have to excuse me, for some of the videos seem to be jerky but werent before compiled into the larger one, they are 720x480 and you can view one Here

but here is the entire thing:

Thor's Intelligence

Statistics: Posted by hivemind — Sun Jul 30, 2006 2:49 pm

A day in the life of a robot

2006-07-30T04:41:32+01:00

Sorry, I should have been more specific. I just bunged the code in, ran it a few times, posted my comment and had to run away

I run my flip directly on hardwood floors. With my own first cut, it works around 70% of the time so it needs some refinement. Where yours fails to work for my RN (and mine fails 30% of the time) is at the obvious point between dropping to his hands and having that momentum carry him over into the forward roll (or not), at that crucial 'DELAY' point. Judging by how fine a line that point is for the RN during the development of this move, I can see how small differences in hand length, zero settings, etc. can cause the move to fail quite easily. I'm going to poke around with it a bit more, see if I can get it towards the level of repeatability of the other stock but complex moves.

But, thanks again for the move! My former-gymnast other half loves it. Never have I had so much acceptance for something so, well, expensive and pointless in her mind We might actually work on a RN gymnastics routine together and put a video together!

Statistics: Posted by beermat — Sun Jul 30, 2006 4:41 am

A day in the life of a robot

2006-07-30T02:32:50+01:00

well im not sure what you mean by "he does not complete the flip" but it could be a zero difference too, i know i have my RN-1 zerod very strangly, so who knows, but if he just starts and moves on too quickly its probably the hands or not a long enough delay.

keep in mind i use a special mat and cement floor for best performance, that may also be your issue.

Goodluck

Statistics: Posted by hivemind — Sun Jul 30, 2006 2:32 am

A day in the life of a robot

2006-07-30T02:28:44+01:00

Actually, just tried yours, hivemind. He doesn't complete the flip. I have stock hands; you have the grippers I believe? Could that be the difference because of the size and weight difference from two extra servos? I have the extra head servo, though.

I have to watch a movie with the missus now; I'll finish up what I was working on for stock hands after, and maybe post that?

Statistics: Posted by beermat — Sun Jul 30, 2006 2:28 am

A day in the life of a robot

2006-07-30T02:19:25+01:00

Thanks for the code! In the event you weren't going to post it, I've been studying the video and have just finished recreating it. Almost. Ugh! I've done the handspring part and the landing prone part, but not the 'stand up at the end' bit. But the hard part is now done. My biggest sticking point (which I just solved prior to looking at your post!) was exactly the same requirement for a DELAY when landing on his hands to let him rock forward. 100 did it for me.

Anyway, thanks for the inspiration, the code (which I will compare to what I came up with from your vid) and the info on the eyes! PWM was the way I was going to go. All I need is a source for the parts. Any clues? I scanned my Digikey catalogue, but didn't see any?

This is the (unfinished standing up part) code I wrote:

'Recreation of the 'flip' move (really a front handspring) as originally demonstrated by hivemind

'Arms out, lean back a tad
SPEED 8
MOVE24 101, 65, 133, 111, 97, 100, 10, 190, 122, 100, 100, 101, 10, 186, 130, 100, 100, 100,101, 67, 132, 110, 98, 100
WAIT

'Rock ankles back, body forward - lands on hands and should tip forward
HIGHSPEED SETON
SPEED 15
MOVE24 101, 128, 187, 167, 97, 100, , , , , , , , , , , , ,102, 132, 188, 159, 98, 100
WAIT

'Small pause to complete rocking
DELAY 150

'Swing legs almost straight with body which is now almost vertical upside down
MOVE G6A, 97, 101, 154, 109, 97, 100
MOVE G6D,102, 106, 160, 104, 102, 100
WAIT

'Landing posture
MOVE24 99, 142, 23, 82, 96, 100,178, 181, 119, 100, 100, 102,173, 180, 112, 100, 100, 100,102, 145, 24, 81, 99, 100
WAIT

HIGHSPEED SETOFF
SPEED 8

'Now get up off your arse!
'TODO
RETURN

Statistics: Posted by beermat — Sun Jul 30, 2006 2:19 am

A day in the life of a robot

2006-07-30T14:52:26+01:00

Well folks, here is the long awaited code, ready to go, all you have to do is compile exactly what i post and send it over, it can work as its own independant program.

Before you you test this though, there are several things you should know. I have modified my RN-1 so that it has three extra servos that could dramatically alter the effects of this flip, the head servo (located in the chest as i explained earlier) and Matt's two gripping claws. Also i have used several rubberized mats on a cement floor (cement reduces inconsistencies and the rubber mat is to make sure it is not damaged) I have noticed that he will work on other surfaces (hard wood and even my carpet) but the cement (or tile floors) with a rubber is the best.

And one final thing: any damages that may incur due to this routine or any variation you may create/come upon are not my responsibility. You should be careful with your robot and use it wisely, give it plenty of space and dont let it fall and there SHOULD be no problems, but i suggest a rubber mat for protection. I hold no responsibilty for any damages as i have tested and re-tested my routine to perfect it. You use this code at your own risk, but i would appreciate that if you were to post it elsewhere you do not claim it as your own, and even if you do, be sure to keep my warnings.

here you are:

'== motor diretion setting ======================
DIR G6A,1,0,0,1,0,0
DIR G6B,1,1,1,1,1,1
DIR G6C,0,0,0,0,0,0
DIR G6D,0,1,1,0,1,0

PTP SETON
PTP ALLON

'== motor start position read ===================
GETMOTORSET G24,1,1,1,1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,1,1,1,1,1,0
'== motor power on ========================
SPEED 5
MOTOR G24

DELAY 5000

'==standard pose==
SPEED 10
MOVE G6A, 100, 76, 145, 93, 100, 100
MOVE G6D, 100, 76, 145, 93, 100, 100
MOVE G6B, 100, 30, 80, 100, 100, 100
MOVE G6C, 100, 30, 80, 100, 100, 100
WAIT
'=================

PWM 2,25

'==wait for user to move away==
DELAY 2500
'=====================

'==lean back and arms out forward==
SPEED 10
MOVE G6A, 100, 73, 145, 83, 100
MOVE G6D, 100, 73, 145, 83, 100
MOVE G6B, 11, 188, 98
MOVE G6C, 11, 188, 98
WAIT
'========================

'==ankle tilt and then "jump" to arms==
SPEED 15
HIGHSPEED SETON
MOVE G6A, 109, 102, 185, 158, 99
MOVE G6D, 109, 102, 185, 158, 99
MOVE G6B, 10, 185, 104
MOVE G6C, 10, 185, 104
WAIT
'==========================

'==wait for momentum to kick in==
DELAY 150
'======================

'=======================================
'=lands on hands and feet:
MOVE G6A, 99, 103, 101, 10, 98
MOVE G6D, 99, 103, 101, 10, 98
MOVE G6B, 190, 185, 104
MOVE G6C, 190, 185, 104
WAIT
'=======================================

HIGHSPEED SETOFF
SPEED 10

'==bring legs underneath==
MOVE G6A,100, 165, 70, 15, 100, 100
MOVE G6D,100, 165, 70, 15, 100, 100
'==================

'==reposition==
MOVE G6A,100, 165, 40, 100, 100, 120
MOVE G6D,100, 165, 40, 100, 100, 120
WAIT
'==========

DELAY 100

SPEED 8
'==standard_pose======================
MOVE G6A, 100, 76, 145, 93, 100, 100
MOVE G6D, 100, 76, 145, 93, 100, 100
MOVE G6B, 100, 30, 80, 100, 100, 100
MOVE G6C, 100, 30, 80, 100, 100, 100
WAIT
'==================================

PWM 2,0
PWM 0,25

SPEED 14
'==arms up==
MOVE G6A, 100, 76, 145, 93, 100, 100
MOVE G6D, 100, 76, 145, 93, 100, 100
MOVE G6B, 100, 160, 85, 100, 100, 100
MOVE G6C, 100, 160, 85, 100, 100, 100
WAIT
'=========

^END OF ROUTINE: As for tempusmaster, thank you for posting my video on your site i will have a follow-up very shortly if you would like to add it to that same page (heh)

and beermat, as for the tr-colour LEDS that im using for the eyes, i am using the PWM ports (as you can see in my code above) instead of the LED ports as i can control the power of each LED

Statistics: Posted by hivemind — Sun Jul 30, 2006 12:31 am

A day in the life of a robot

2006-07-29T21:56:54+01:00

Excellent flip, very impressive! I'd like to join the eager queue of owners salivating over the chance of you posting all your hard-earned MOVEs on this forum.....

Also, could you talk a bit more about the eyes mod? It looks like you have tri-colour LEDs in there? I've been trying to source 5v 3mm tri-colour LEDs, but not had much success. Also, how do you control the different colours?

I too added the servo for the head movement, and it's a really tight squeeze, as you said. Hitec needed to add a couple more mm to make a standard RN servo fit, IMHO.

Statistics: Posted by beermat — Sat Jul 29, 2006 9:56 pm

A day in the life of a robot

2006-07-29T20:34:39+01:00

Hivemind,

That flip is awesome, totally amazing, I never thought I'd see a robonova doing something like that - hats off to you

Pev

Statistics: Posted by Pev — Sat Jul 29, 2006 8:34 pm