Page 165 - Robots Androids and Animatrons : 12 Incredible Projects You Can Build

operation of the speech-recognition program slows the operation
of any other program that runs using voice recognition.
There are many applications to voice recognition aside from robot-
ics. Speech recognition will become the method of choice for con-
trolling robots, virtual reality (VR), appliances, toys, tools, and
computers. Because of the far-reaching potential of this technology,
companies are developing speech recognition. The ability to control
and command a computer (or appliance) by speaking directly to it
will make working with that device easier, and more efficient and
effective. At its most basic level, a speech-controlled device allows
the user to perform parallel tasks, (i.e., hands and eyes can be
busy elsewhere, while continuing to work with the computer or
appliance).

There are three construction projects outlined in this chapter.
The first project is a speech-recognition circuit. The second pro-
ject interfaces the speech-recognition circuit to a mobile platform
[radio-controlled (R/C) car]. The third project is a general interfacing
board for the speech-recognition kit.

Project 1: Programmable speech-recognition circuit
144
The first project is a programmable speech-recognition circuit. It
is “programmable” in the sense that you can program the circuit
to recognize up to 40 unique words of your own choosing. The
heart of the circuit is a single integrated circuit (IC), the HM2007
speech-recognition chip. The chip provides the options of recog-
nizing either 0.96-second (s) or 1.92-s word lengths.
Using 0.96-s word lengths enables the chip to recognize 40 indepen-
dent words using an 8K 8 static random access memory (RAM).
You have the option to switch to the longer 1.92-s word length. While
this reduces the word recognition count to 20 words, the longer
word length can be used for phrases instead of isolated words. The
circuit we will build here will use the 0.96-s word length, 40-word
recognition library.

Learning to listen

We take our speech-recognition abilities for granted. Listening to
one person speak among several at a party is beyond the capabilities
of today’s speech-recognition systems. Speech-recognition systems
like ours have a hard time separating and filtering out extraneous
noise.

Team LRN
Chapter seven