Hi,
I need some help with a project of mine, i'm using speech to control a robot and i'm planning on using a neural net to do the speech recognition and if anyone has any code that could help i would be most grateful. The idea is that the speech signal goes into the input is processed by the "hidden layer" and then comes out of one of the 6 oupputs that could then be connected to the motor control circuits and if someone says forward the neural net processes the signal and activates the right output. The command are forward, reverse, left, right and stop to which each would have a output and another output for all other words? Like i said, any code would be great. Thanks

My assumption is that your system will receive as input a sampled, digitized audio waveform. Is that correct? I also wonder: are you more interested in solving the speech recognition problem or using a neural network? I imagine that this problem, with its small vocabulary requirement, could be solved relatively easily without using a neural network.

-Will Dwinnell
http://will.dwinnell.com