king spider does not what it is supposed to do

[ID_Stephan]

Hy to the community,

since this is my first post, I want to introduce myself shortly.

Identification: Stephan
Occupation: student of industrial design ,right before graduation
Age: 27

Last year I took part in an internship for a japanese company. Which was my entry in robotics. Well, the internship was not, but my time in Japan.
First I got infected by all the mecha and anime stuff they have, and now my graduation project has become a robot design aswell.

I got myself the bioloid comp kit.
I have no clue about programming and C.

For my project I needed a hexapod frame to style some housing for.
So I thought the advcanced spider of the bioloid kt would be perfect.

I hoped just to build the kit. ...feed the cmd5 with the example data and thats it.

Assembly check of my spider is positive and I uploaded the example behavior file aswell as the motion file to the cmd.

When I clap hands it "awakens" as shown in the expample vid, then it goes on walking, and after some time changes direction. ...strangly it feels like the direction change move is not initiated by any sensor feedback, since it changes direction even if nothing is in the way...sometimes it walks against an obstacle and doesnt change direction.

First I thought my sensor dynamixel was damaged, so I tried building a minor complex robot like the flower and tested all the sensor funtions with it.
ALL POSITIVE!

I have no clue what might be the problem.

Really hope somebody here can help me out, or give me a hint.

Maybe somebody can send me "spider files" which are sure to work.


BTW. I builded the humanoid, too. ...everything was working, except the walk-mode.

Are those example files somehow "unready" ?


really hope my english is understandable, I*m a little bit out of practice.
greetz, Stephan

[ID_Stephan]

I just downloaded the new software from robotis.

Reinstalled it. and then

1st changed robot file in behaviour control programm

2nd uploaded example motion data

3rd opened and uploaded example behaviour file.

Still the same.

I also watched the vid again and my spider just starts walking when I clap.

In the vid the spider gets up on the first clap and starts walking on the second sound input. .


I do not get where the problem might be.

checked the behavior file in the program... it seems logic and alright to me so I did not change anything.

[siempre.aprendiendo]

It works, Stephan.

Have you checked that all your AX12+ are properly mounted and centered at the middle point, position 512?

Sometime I have suffered some strange problems when an AX12+ has been wrongly mounted or their initial position isn't at the middle point.

Hope it helps

[ID_Stephan]

Many thancks for replying,

but I assembled everything correctly.

I checked the positioning lines on the servos when building the legs and made sure the direction is right.


When I try to execute the motion pages in the motion editor program by entering the page number and execute play, they all work like discribed in the text file.

This is proof to me the assembly of the robot has no mistakes.
If I made mistake in building and/or the initial position is not right the motion pages would not execute correctly in the motion editor and my robot would do weird movements or even damage itself, right?
Assembly check is positive, too.

I guess the problem might be the sensor.

The spider gets up when I clap and executes motion Nr.
50 Behavior1 Demo scenario(for normal state)
like it is supposed to be in the Behavior Control file.

But thats it ... it walks some time, and after a while it turns.

No matter what I put in its way, hand/book/wall, it never executes another
motion.

I tried to modify the numbers of the sensor input in BCP but no effect.

I tried different robots (less advance) and checked if sensor might be damaged, but it seems to work for the other bots.

For the spider config, I even tried changing execution of the obstacle motion by clapping twice. No result either.

Somehow it seems to me that it gets stuck in the motion.


I have no clue. I am totally confused.

[AlphA]

Have you looked at the behavior control program or the motion data?

While in the "normal state loop", the spider is programmed to perform motion 50 (demo scenario). This is not to simply walk forward until it detects an object. If you look at the motion data, you'll see that motion 50 loops a walk cycle a few times, then turns right. I don't have my CM-5 with me, so I can't confirm exactly what motion 50 is doing. But, I'm pretty sure that's how it works.

So, if your bot stops walking forward, BACKS UP then turns, it is performing its "obstacle response". On the other hand, if your bot simply goes from walking forward to turning right (without backing up), it is carrying out the demo scenario, which is normal.

If it fails to detect obstacles, check your lighting. Bright lights can sometimes confuse the sensors. If I was you, I would throw together a simple program telling the CM-5 to print AX-S1 sensor data. Then, move your hand near each sensor and make sure the data output corresponds.

In behavioral control programmer, replace motion 50 (demo scenario) with motion 2 or motion 3 (forward walks) and you should should have the behavior you want.

[siempre.aprendiendo]

Ummm, I remember now that there are some interferences between the sun light and infrared. IR detects as if there were an obstacle and "avoid" it. I had this problem when the spider was near a big window

[ID_Stephan]

First I want to say a big THANCK YOU for your support.

But my problem still is not solved.

The Problem could be located in the Behavior control file.


The motions are executed correctly when using the motion editor.


I tried to change the initiation for the obstacle response to a sound input (2claps)...still not working.

It seems like the bot is stuck in the first loop it executes and does not have any interests in further input.



Different approach:
When I change the first motion from motion 50 to motion 2 ist makes 1 step and then stands still ... then I can give it further input like IR making it sit down.

but then when the attack move is initiated it get stuck in the loop again.

I apologize, I do not have any knowledge about programming at all.
So I can not "read out any data"
I checked the functions of the sensor with an "easy" bot (the flower) every sensor function seemed to work.

[trey3670]

did you d-load just the mtn file or both the mtn and bpg?

[ID_Stephan]

I downloaded both.

First the motion file then the behavior file.

I can assure that my problem is not covered by mistakes that can be solved with the instruction book.

[JonHylands]

You might have a bad AX-S1... I would write a little program that tests it on its own - connect just the AX-S1 to the CM5, and test each of the three IR sensors to see that they are working correctly.

- Jon

[ID_Stephan]

Yes, my guess is, too, that the Sensor module is damaged somehow.

Unfortunatly I am not able to write a program like the one you advised. I have absolutly no knowledge in programming.

My approach to robotics came from the field of design.

I am a designer, not an engineer or a programmer, ...too bad for me.

I just wanted to have the spider do what it is supposed to by its example programm. ...didn't expect any problems like this.

Guess robots are NOT less complicated then any woman

[SK]

Well seeing you´re from Darmstadt like me (the world is small ) I can check if you can pay a short visit to a certain robotics lab I know where there are unused AX-S1 sitting around that you could use to test your spider. PM me if you´re interested.

[AlphA]

You can use mine.
Bioloid IR Sensor monitor.
You may want to start tinkering around in the behavior control programmer. It's designed for non-programmers to use. It is a valuable skill for troubleshooting and tweaking. Otherwise, you will probably get stuck on little issues over and over again.