Programming the PLC

A PLC project must be programmed in order to use TwinCAT Speech. For a quick start, this is explained below on the basis of this sample.

1. Create a new PLC project.
Programming the PLC 1:

2. Add the Tc3_Speech library.
Programming the PLC 2:

3. Insert the following code from Sample01 into MainASR.

Declaration part:

PROGRAM MainASR
VAR
// ASR Trigger Variables
bListenTrigger        : BOOL     := FALSE;
bListenStopTrigger    : BOOL     := FALSE;
// TTS Command Configuration
nConfigIdTTS          : UINT     := 200;
// ASR Command Configuration
nConfigIdASR          : UINT     := 100;
// Language Id for TTS Output
nLanguageId           : UINT     := 1033;
// fbTTS is required for providing TTS responses upon successful recognitions in this sample.
fbTTS : FB_TextToSpeech     := (nConfigurationId := nConfigIdTTS); 
fbASR : FB_SpeechRecognition:= (nConfigurationId := nConfigIdASR);
fbRetrieveUtterance : FB_RetrieveUtterance;
// Update Code before Release
// TTS Variables (required for TTS responses upon recognition)
bSpeak                : BOOL     := FALSE;
{attribute 'TcEncoding':= 'UTF-8'}
sText2Speech : STRING(4096) := '<speak> TcSpeech beta demo project is greeting you </speak>';
// ASR Variables 
bListen               : BOOL     := FALSE;
nLastRecoId           : ULINT    := 0;
sLastRecognition      : STRING(4096)     := '';
fConfidenceThreshold  : REAL     := 0.7; // TcSpeech will inform PLC if recognition confidence is larger than this threshold
bInit                 : BOOL     := FALSE;

Process part:

IF NOT binit THEN
    fbTTS.SetAmsNetAddr(GVL_SpeechDemo.sAmsNetId);
    fbASR.SetAmsNetAddr(GVL_SpeechDemo.sAmsNetId);
    binit := TRUE;
END_IF
// ASR - Automatic Speech Recognition
// Trigger Start/Stop Listening
IF bListenTrigger THEN
    bListenTrigger:=FALSE;
    bListen := TRUE;
END_IF
IF bSpeak THEN // for answering a recognized ASR command ... 
    fbTTS(sUtterance := sText2Speech,bSpeak := TRUE, nConfigurationId:= nConfigIdTTS, nLanguageId:=nLanguageId);
    IF NOT fbTTS.bBusy THEN
        fbTTS(sUtterance := sText2Speech,bSpeak := FALSE, nConfigurationId:= nConfigIdTTS);
        bSpeak := FALSE;
    END_IF
END_IF
IF bListen THEN
    IF bListenStopTrigger THEN
        fbASR(bListen := FALSE,nConfigurationId:= nConfigIdASR);
    ELSE
        fbASR(bListen := TRUE,nConfigurationId:= nConfigIdASR);
    END_IF
    IF NOT fbASR.bBusy THEN
        fbASR(bListen:= FALSE,nConfigurationId:= nConfigIdASR);
        blisten := FALSE;
        bListenStopTrigger := FALSE;
    END_IF
END_IF
// Check if new Recognition is available
IF nLastRecoId <> fbASR.nRecognitionId THEN
    IF NOT bSpeak THEN
        IF fbAsr.fRecognitionConfidence > fConfidenceThreshold THEN // if recognition better than treshhold, answer via TTS. Next cycle will process via FB_SwitchLanguage
        // Check if Recognition Confidence is high enough
        nLastRecoId := fbASR.nRecognitionId;
        fbRetrieveUtterance(nLanguageId := nLanguageId, sRecognitionTag := fbASR.sRecognitionTag,fRecognitionConfidence := fbAsr.fRecognitionConfidence, sText2Speech => sText2Speech);
        bSpeak := TRUE;
        END_IF
    END_IF
END_IF
END_VAR

4. Set "bListen" to TRUE to make a speech input.
The TwinCAT configuration must have been activated beforehand.

The identification number for the ASR configuration to be used is present in the code as ConfigIdASR.