Return to site

Voice Recognition Network Module

All-in-one package for voice recognition in home appliances

Unlike products such as smart speakers, where speech recognition is considered at the planning stage, it is not easy to implement speech recognition functions in existing home appliances.
This is because not only understanding of speech recognition system, microphone input and pre-processing, speech recognition engine, but also network connection, communication with main system of product should be considered.

Manufacturers will be very welcome if there are any hardware modules that solve these issues. This is because the manufacturer does not have to worry about other issues besides designing and assembling the product considering the location of the module.

MightyWorks introduces the Voice Recognition Network Module(VRNM) for home appliances.

The VRNM consists of a microphone input part, a voice processing part, a trigger(command) recognition part, and a network part.

  • The microphone input part is composed of one or two microphones, and is the foremost receiving part of the user's voice command.
  • The voice processing part processes the voice signal inputted through the microphone, by the pre-processing algorithms such as beamforming, noise suppression, echo cancellation and so on. The processor can be Cortex-m4 or Cortex A series.
  • Trigger recognition part serves to wake up the product before the user commands or questions the product by voice. Calling a trigger allows you to ask or command the product through natural language. Or you can say simple commands that are already ported to the product so that you can execute commands from the product itself without having to connect to the cloud through an external network.
  • The network part is responsible for connecting to the cloud speech recognition engine over Wi-Fi. It is a network that is required when users send commands or questions spoken in natural language to the cloud engine and receive the results. When it is not possible to connect directly via wi-fi, it can be connected to external device via Bluetooth connection, and this external device can take charge of cloud communication via internet. The voice recognition result received through the network unit is sent to the main processing unit of the product, where it answers the question asked or performs the function requested by the user. If the main system of the product has a network, the speech recognition module can remove the network part and leave this role in the network part of the main system.
MightyWorks' speech recognition hardware module basically provides its own voice signal processing technology and trigger engine. The cloud engine can be supported by Alexa, Google etc. according to the requirements of the manufacturer or service company.
Processor and memory, wi-fi and Bluetooth are selectable to the needs of the manufacturer.

All Posts
×

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!

OKSubscriptions powered by Strikingly