Spectra voice box and the included vicCONTROL voice dialog system offer an internet-independent, local voice control, which allows the user a very flexible way of voice input. Several keywords (intents, slots) and parameters (values) can be embedded in any phrase, without limiting the order of the keywords and the structure of the commands (Natural Speech Understanding - NLU). A semantic evaluation interprets the language inputs, determines the task to be performed and the parameters to be controlled. The universal IoT protocol MQTT enables reliable coupling and independent spatial distribution of voice operation and machine control in the local IP network. Due to the availability in 30 languages, worldwide use is no problem.
The web-based development tool vicSDC enables the industrial user of vicCONTROL to independently create and extend his speech dialogs during the design phase by entering examples and marking intents, slots and values.
Customer-specific voice control consists of the application-independent, generic vicCONTROL voice dialog application and application-specific voice resources (voice vocabulary with intents, slots, values, voice output and control interface to the target application), which are created by the user during this design phase.
The Spectra VoiceBox is delivered with the generic vicCONTROL application and a license for using the vicSDC web tool.
To use the voice control, the first step is to define the desired national language and to select a keyword (e.g. "Hello VICki").
The definition of control tasks (intents) and their value assignment (slots) are defined by the user examples and the keywords and slots are marked manually.
The speech dialog is completely created in the web-based development tool vicSDC. All resources such as example sentences, intents, slots and speech outputs as well as control interfaces are created and managed in one project, so that subsequent adjustments can be made easily.
The created project is compiled using AI-based methods, checked for errors and stored in a resource package.
The compilation process adds variations of the natural language.
The created resource package is made available to the user in the web tool vicSDC.
The resource package is transferred to the Spectra VoiceBox embedded system or the Spectra PowerTwin Panel PC. There it is the basis for the customer-specific voice application.
The user has the possibility to use the learned commands in a natural language. Keywords are used to activate the speech dialog only when required. As soon as the speech recognizer recognizes a command via the microphone, it is processed further.
The voice commands are linked to an MQTT command, which is either processed locally by the machine controller or transmitted to the machine via an existing ethernet interface.
For our company´s day of a "open house" we presented something very special to our visitors: a "Magic Mirror" with a voice control from Amazon Alexa and a Raspberry Pi. Read more about it in our construction manual.
Everyone knows the situation - at door intercoms or in public transport announcements, speakers are sometimes difficult to understand. Several factors can be responsible for this.
What parameters are these? How can speech intelligibility be evaluated using measurement technology and how can it be optimised?
Internet-independent voice control and a freely configurable voice dialog - the ideal combination for a wide range of applications in industry and automation.
+49 351 40752650