Chippu has easily learnt to recognize human speech. It uses Julius for speech recognition as explained in the previous post. The following video shows the speech recognition abilities of chippu.
As you can see, it works flawlessly. The complete source is available at the svn
The rxgraph output of the system looks like this:
As it clearly indicates, the node named “talker” takes care of speech recognition and outputs the recognized speech as a string message named “speech”. The “voice_nav” node will then produce the control signals based on the messages of the topic “speech”, and publishes the control signals as a topic named “drive” to the “serial_node” which interfaces the Arduino.