mtiberia
Canada
Asked
— Edited
Since DJ has fixed the bug with the UART, thanks again DJ, I was able to use EZ-B to write a script showing how you can program a motion for your robot through " demonstration " .
There are a few people out there experimenting with this method of programming their robots .
One interesting example uses something called " goal babbling " , where the operator demonstrates different moves to reach a goal, The robot then tries to optimize the motion and even learns to adapt when an object is placed in the way of it's goal.
Really cool stuff , and now I can use EZ-B to replicate some great work that is going on out there in my own house !
Keep up the good work DJ !
I want to know more! Please expand on this programming idea.
This is a good start using the NAO robot.
DemonstrationlearningonNAO.pdf
I'll try that again.
NAODemonstrate.pdf
A perfect example, now if I can only teach my robot how to overcome these pesky little bugs !
I looked at the "Demonstration-Guided Motion Planning" paper.
Now if I can only teach my robot calculus, I could program a motion for my robot through " demonstration " .
I do thank you for the reply, but I still have no idea how to implement demonstration programming (or machine learning) for motion controls.
It is a great goal.