Building on Justin's thread on creating a brain. There are many ways to create artificial intelligence and this is just what I have come up with using the tools ARC has currently available.

I have designed my project to be Service/Agent based. When the Robot is powered up the connection script launches the init script.

ControlCommand("Script Manager", ScriptStart, "Init")

This initializes all variables and launches the Voice Service and the Auditory Service and unpauses the Speech Recognition plugin. The only Speech Recognition phrase I have entered here is the robot's name and the is UNAS (meaning First One) Stargate SG1 fans will have known that. And that script looks like this:

ControlCommand("Speech Recognition", PauseOn)
ControlCommand("Bing Speech Recognition", StartListening)

As you can see when the Robot hears his name "UNAS", it pauses the speech recognition module and starts Bing Speech listening.

Here is the Bing Speech Script

ControlCommand("Bing Speech Recognition", StopListening)
WaitFor($IsListening = 1)
ControlCommand("Speech Recognition", PauseOff)

As again you can see after the speech has been captured the script stops Bing speech listening and unPauses Speech Recognition.

The WaitFor($IsListening = 1) we will see later in other scripts for now we can just ignore that line.

Now to the fun part:

The Auditory Service's first line is waiting for the $BingSpeech variable to change and when it does that is when the magic happens.

:AuditoryService
#The Auditory service monitors the $BingSpeech variable for speech and then routes this speech to
#the proper Agent for processing.
#
#***************************************
WaitforChange($BingSpeech)
$HumanSaid = $BingSpeech
$BingSpeech = ""
$IsListening = 0
#***************************************
#Convert the first letter of the captured speech to lower case
$Trash = $HumanSaid
ControlCommand("Agents",ScriptStartWait,"ConvertToLowercase")
$HumanSaid = $Trash
#***************************************

When the $BingSpeech variable changes, its contents are saved in the $HumanSaid variable and the $BingSpeech variable is cleared. Here is where we see the $IsListening variable set to 0 letting the other scripts know that the robot is currently processing and not listening.

The next few lines converts the first letter of the spoken speech to lowercase, you will see why later.

Next the script goes thru a bunch of if statements.

#BlankCanBlank
if(Contains($HumanSaid," can ") & !Contains($HumanSaid,"What can "))
  ControlCommand("Agents",ScriptStartWait,"BlankCanBlank")
#CanBlankBlank
elseif(Contains($HumanSaid,"can ") & !Contains($HumanSaid,"What can ") & !Contains($HumanSaid," can "))
  ControlCommand("Agents",ScriptStartWait,"CanBlankBlank")
#BlankHaveBlank
elseif(Contains($HumanSaid,"have") & !Contains($HumanSaid,"Do "))
  ControlCommand("Agents",ScriptStartWait,"BlankHaveBlank")
#BlankIsInBlank
elseif(Contains($HumanSaid,"is in"))  
  ControlCommand("Agents",ScriptStartWait,"BlankIsInBlank")
#BlankIsThePluralOfBlank
elseif(Contains($HumanSaid,"is the plural of") & !Contains($HumanSaid,"what ")) 
  ControlCommand("Agents",ScriptStartWait,"BlankIsThePluralOfBlank")

I did not include the entire script just enough for a good example So the if statements look at the speech phrase for specific words in specific orders and calls the corresponding Agent. For example we will use the following speech phrase "dogs can run" The if statements will match this and call the CanBlankBlank Agent

:BlankCanBlank
  $IsListening = 0
  $Junction = "can"
  $SpeechLength = Length($HumanSaid)
  $JuncIndex = IndexOf($HumanSaid,$Junction)
  $FirstPart = SubString($HumanSaid,0,$JuncIndex-1)
  $LastPart = SubString($HumanSaid, $JuncIndex + Length($Junction) + 1,$SpeechLength - ($JuncIndex + Length($Junction))-1)
  $SearchPhrase = $FirstPart + "_:::_" + $LastPart
  $MemoryType = "BlankCanBlank"
  ControlCommand("Agents",ScriptStartWait,"WriteMemory")
  if($WriteMemoryCheck = 0)
    $SayWhat = "I knew that!"
  else 
    $SayWhat = "I now know that " + $FirstPart + " " + $Junction + " " + $LastPart
  endif
$IsListening = 1

The Agent Parses the captured speech takes what is before the word can and stores that in $FirstPart and what comes after the word can and stores that in the variable $LastPart Then loads the variable $SearchPhrase and $MemoryType then calls the Write Memory Agent.

This Agent opens the file and searches the file for the SearchPhrase. I will go into detail of the read and write memory Agents in the next exciting episode of Creating a Simple Brain.