This post is merely to keep track of sources I find, which could be useful soon in the project.
The paper above discuses some problems with batch training of neural networks and tries to argue that on-line training has more advantages, especially for larger training sets.
A blog that explains the backpropagation algorithm using matrix concepts to compute the outputs and error signals.
It is a blog that explains neural networks in a simple way. It is easy to understand and concepts are explained incrementally. It was very useful in understanding the backpropagation.
Some simple slides explaining basic concepts of how quad-copters work.
A more detailed description of how a quadcopter operates. It includes some physics concepts.
This blog explains how the PID algorithms work with a quadcopter. Additionally, there is a java applet that lets you compare the performance of different PID constant values, or the performance of the quadcopter under a neural network control.
A paper that describes the implementation of a neural network as a flight controller for a quadcopter.
Background Subtraction Papers