Great stuff @utrinque , that is crazy cool, I liked it a lot. I especially liked the build by flying from the link. where art meets quad coolness
Simon

 

Pretty cool technology

 

I loved the balance stick on the first video. Great programming and some sophisticated electronics there. simply amazing....

 

I like this
Simon

 

I've got an algorithm for that somewhere, it's based on neural networks and fuzzy logic, I'll scan it for you if you want (and I can find it) - I think it's in a book by Timothy Masters?
Simon

 

Oh really! I am interested!.. I want to create something in Java that models that physics - just for fun....If you can scan it i would be very interested @simon

 

I'll see if I can dig it up tomorrow - I think it has the code in C++ (or at least code snippets of the bits you need) too if that helps you
Simon

 

That was amazing @simon... If only we could find an energy source to keep the little suckers in the air for a much longer period of time. Failing that, there could be landing/recharge stations... ie pick up block, secure block land on recharge station. Imagine the possibilities! If the quad copters were strong enough to pick up a brick and another apply mortar or use a different method for securing then with a clerver bit of programing, houses could be built much quicker! Not only that it might eliminate the need for scaffolding in very difficult to reach places!

Re the code, C++ is very similar to java OOP so that would be cool as well...

 

We thank @utrinque for that, it was on the name link in the OP. Cool research
Simon

 

Software, algorythms seem to be more crucial than hardware in future applications.

Actually he uses relatively cheap RC components and widely available cheap electronics.

Are we the fossil in the world where cars do not need drivers, planes and helis have no pilots? Chips can fly them much better than humans. Deep learning techniques (including neural networks) will be able to create new 3D smack tricks no human can execute etc.

Quad with only two working propellers is able to move in all directions... controlled by algorythm. No human can do this.

FBL gyro with GPS should be able to save model heli performing autorotation in case we loose tail control.

It needs altitude, RPM, descent rate and ground distance. If your field is flat... you can have all those parameters from already existing sensors. Processor should only memorise GPS altitude at motor start before take off.