Has anyone (Arcbotics included) had any luck coding any sort of program to let Hexy run on its own, maybe walk until near something and then turn around, repeat? Is that even possible with the controller board, or are we looking at eternally needing to be “connected” (even if via bluetooth)?
Hi - I am in the process of attempting something like this but no luck yet. The only comment I would have is that the the processor on the controller board isn’t powerful enough to run any kind of decent avoidance and movement algorithm. I think it is working hard enough to maintain accurate servo timing and to service the serial port. I am planning on keeping it connected via bluetooth and its ‘brain’ running on a good old fashioned powerful x86. Perhaps though a basic forward and turn algorithm could run on the controller…
John.
[quote=“ss023459”]Hi - I am in the process of attempting something like this but no luck yet. The only comment I would have is that the the processor on the controller board isn’t powerful enough to run any kind of decent avoidance and movement algorithm. I think it is working hard enough to maintain accurate servo timing and to service the serial port. I am planning on keeping it connected via bluetooth and its ‘brain’ running on a good old fashioned powerful x86. Perhaps though a basic forward and turn algorithm could run on the controller…
John.[/quote]
I wonder if a raspberry pi would be too heavy to use a the “brain”? Runs on 5volt already present on Hexy.
Oooh I hope not. I’ve already got plans to make a hexy + pi + wiimote controlled robot. But with the extra gear I’m wondering if hexy can handle it, especially with all these reports of broken servos.
Does anyone know the carrying capacity of hexy?
We’ve carried up to 1lb on it, walking slowly. I imagine you could walk better with more weight if you used a wave gait (only picking up one leg at a time).
So I guess the 1.6oz Raspberry Pi won’t break hexy’s back.
(…quickly does the math…)
So I wonder if you could add 10 Raspberry Pi’s and have the first walking hexapod computer cluster…
evil genius laughter
[quote=“distantcam”]So I guess the 1.6oz Raspberry Pi won’t break hexy’s back.
(…quickly does the math…)
So I wonder if you could add 10 Raspberry Pi’s and have the first walking hexapod computer cluster…
evil genius laughter[/quote]
Sure wish the forum had a mod so I could +1 your post… 
I agree with ss023459 the controller lacks the power to perform complex logic. I am working in implementing a firmware that accepts movement commands over bluetooth from an inverse kinematics algorithm implemented in java. I have ordered an accelerometer/magnetometer from Adafruit and hope to communicate that information as well as information from the PING))) sensor back over the BT link to the mothership. I think the absolute rotation reference from the magnetometer will allow a map of the room to be generated using the PING))) distance data and it’s servo’s position.
My first design goal is to get the Hexy to autonomously map a room using the PING))) data; we’ll see how reasonable that is soon.
I wish you luck. Be aware that the magnetometer will be fairly inaccurate as you wander around, as anything that generates an electric field tends to mess those up (like say, the current running through the walls).
Servotor 2.0 should be out soon, and this should help autonomy a lot. We’ve crammed all the extra stuff needed to run the servos into a library, are writing a better driver for the ultrasonic, and some other fun stuff, too.
Do you mean a new Servotor board with more memory, or a new firmware update for the existing board?
firmware update for the existing board
Do you do development on the Servotor code offline, and only push when you’re ready to release?
I’m currently maintaining my own fork and would like to keep up to date with any changes.
Yeah, code development is mostly offline. I can develop with just a servotor32 and my salae logic almost anywhere, but I test with a hexy before I push code live, to make sure it works as advertised. However, I don’t always carry one with me wherever I code, so thus the offline development.
There’s a technique where you do all development on a branch in git, and when you’re ready to release you move it over to a release branch.
New firmware is posted! Have fun with it. I should have the PoMoCo update in a week or so.