Let's Make Robots!
tc  
AttachmentSize
screenshot of the simulator433.1 KB

All my projects & progress are available at marginallyclever.com.

About

This is version 1 of my hexapod robot project.  Version 2 is now in development.

I started this project March 9, 2010.  Most of my time has been spent writing the simulation software - a 3D model of the robot that I used to perfect the inverse kinematics.  In the early versions the PC did all the kinematics and then told the robot (via serial link) where to move the servos.  Once the simulation was good enough I moved all the kinematics code to the robot.  Now all a controller needs to do is say "move this way" and the robot goes.  I used the cheapest servos I could get and so I'm not certain it could support much more weight.  I'm also unsatisfied with the plastic parts.  Your mileage may vary.

The code for the IK will NOT fit on a duemilanova arduino.  They don't have enough ram.

Instructions & Plans

Full assembly instructions and source are provided at the instructables.com link.

All the code and DXF files for parts can be downloaded from http://visual-hexapod.sourceforge.net/ if you want to run the simulator or cut your own parts and try to assemble it yourself.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Korel's picture

Cool robot my friend, I like it. Keep up the good work :-)

Hi, do you think that the bot could work with an a arduino mega? Where is the pde file for the bot? I could not find it... I am curious about the code and the IK ! Looks like a really interesting project! Federico

aggrav8d's picture

You'll need at least 32k of ram.  Duemilanova has 30k.  Mega has 124k so you should be alright.

I prefer the DisplayDuino & ServoBoard because it's so easy to attach the servo pins and I don't have to do my own PWM.

OddBot's picture

Your statement about doing your own PWM has me confused. Most Arduino boards can drive up to 12 servos in the background using the servo library. Using the writeMicrosecond command will give you much better resolution.

aggrav8d's picture

Try handling 18 servos at once while using all the CPU power to calculate the inverse kinematics.  By offloading the PWM onto the servo board I make it easier to attach the 3-pin servos to the assembly AND I reduce some of the calculation overhead AND I make sure that the pulses are sent when they're supposed to be sent.

I wish to try your code, i could erase some procedure to fit it the 2009 just to see how it works... Where I can find it?

Can you tell us something more about the visual hexapod? I have tried it but without a real robot so I can't still understand it completely.

 

Thanks for your time, Fede

aggrav8d's picture

All the code and DXF files for parts can be downloaded from http://visual-hexapod.sourceforge.net/

BoBBy's picture

I think that legs are too thin, entire body is rocking during walk. Or is it because of your servos, they aren't strong enough?

aggrav8d's picture

The plastic bends and the servo resolution is low - the servoboard only gives 256 positions while the servo can accept upt to 1024.

I do not know which is the cause of the shaking.  If I did I would have fixed it already.  I'm designing a new robot now that should eliminate the jitter.

BoBBy's picture

So maybe try adding brackets glued to the legs. Because i think thats because of not enough rigidity of the structure.