wow, really nice, sturdy
wow, really nice, sturdy looking build and it works pretty good too. congrats 
Love this. Very informative
Love this. Very informative and thank you for your detailed explanations. I very much like the part on PID and am interested in how you are using I2C. Lots of really useful stuff here.
Look forward to see how you flesh this out although its pretty complete already.
Clive
Very good work. I like the
Very good work. I like the idea of cascading the PID (PI in this case) algorithms. I may have to try this method with mine. I love the use of steppers as well. Do you have a number for the overall wheel diameter? I tried to scale it on my screen with the 5mm bore of the hub for the motor shaft and came up with about 130mm. Very original all around. Keep it up and thanks for all the great information here.
HelloThank you for your
Hello
Thank you for your comment. The overall wheel diameter is approximately 105mm.
Thank you all
Thank you for all positive comments, I’m definitely encouraged to give it some new skills.
Very nice work. And I don’t
Very nice work. And I don’t say this often
Nice!
Nice and fresh idea! Carbon frame looks very nice!
nice work
nice work
Very cool to see
Very cool to see someone use the PI for what it can do. You are getting me excited about playing with a PI. I have to finish my current Arduino project, and then onto my RasPI!
Your detailed explanations are excellent. I look forward to seeing what else you can do with this configuration. Thanks for sharing all of your hard work.
Regards,
Bill
Cool Robot
Really great! I appreciate the great explanation and giving key points.
Impressive project
Really impressive. It appears that the balancing is much better than other 2-wheelers.
You wrote " I’m using stepper motors in this robot, no special reason for that."
Do you think, that the steppers a a reason for the perfect balancing?
Thank you for your comment.
Thank you for your comment. I really don’t know if the steppers are the best option. This is my my first 2-wheeler, I’ve seen other good robots (youtube) that use DC motors. The balance process gave me a lot of work and it is not perfect, some aspects need to be improved. Maybe the reason for good results is to work hard
work hard
Yes, may be you are right with the “work hard” statement.
How log have you worked on this robot?
I’m working in this robot
I’m working in this robot since the begining of 2013 (first drawings).
Awesome!
Awesome robot Samuel! Very nice and clean construction and very informative writeup. I love balancing bots and you have combined it with image processing stuff and took it to a whole new level! I think this is the first balancing bot i have ever seen which uses steppers to drive the wheels instead of servos or traditional gearbox motors with encoders etc. Very stable too and one thing, please correct me if i am wrong but i thought analog gyros/accelerometers were slower for such applications, anyways seeing the video it seems they are not!
I completely agree with Bdk and nhbill that this is one of those robots that truly utilizes the real potential of Rpi.
Thanks for sharing this. Looking forward to more progress
Thank you for your positive
Thank you for your positive feedback
Honestly I don’t know if this gyro is slower than the digital ones. I’ve seen a quadcopter controller board using this gyro (kk board I think). Most recent gyros and accelerometers use digital interface and are probably better than this one, because of this what you say makes sense.
Just a short note for anyone
Just a short note for anyone interested - digital gyros and accelerometers are in fact slower because they have to convert the internal sensor signal to digital and then send the data out via serial, rather than just amplifying the raw analog signal and passing that to the output.
Overall you might be reading an analog sensor more slowly though, because the ADC on your microcontroller might not be as fast as the onboard ADC inside the digital sensor.
There’s nothing wrong with analog gryos and accelerometers, but a digital sensor is easier to handle because:
A) the ADC is taken care of internally.
B) the serial connection from the sensor to your microcontroller is much less sensitive to noise and interference.
i didn’t know that…
thanks for the info Telefox
Nice job with the
Nice job with the instructions by Signs
arduino for the balancing part?
hey, i’m thinking about building your bot. however, i want to do the lower part as arduino. i got myself stepper motor shields with motors and a shield with a combined gyro / accelerometer. i could imagine that the kalman filter then is already implemented in that shield? do yo think arduino is fast enough todo the balancing?
secondly, do you have any starting points on the openCV on raspberry pi? did you write the whole code yourself, did you have to compile opencv for raspberry pi yourself?
great inspiring work!