The STMicroelectronics SL-VUI-CLOUD-01 is a cost-effective way to integrate AVS for AWS IoT Services® into smart devices, enabling state-of-the-art voice control based on natural language understanding, so users will enjoy an enhanced experience with targeted IoT end products.

Be able to talk to Amazon Alexa® and control smart home devices, get help, listen to the news, check the weather forecast, play music, and more.

The software package implements the audio front-end, Amazon wake word, audio playback and Amazon Alexa communication protocol software.

The SDK runs only on internal memory, providing maximum integration and a cost-effective solution.

The SL-VUI-CLOUD-01 is built using a modular approach that allows for easy prototyping and debugging, as well as easy adjustment of specific microphone spacing, user interface and audio output requirements.

The solution consists of a motherboard with an STM32H743 microcontroller and certified Wi-Fi module, and a daughterboard with two high-quality MP23DB01HP mems microphones spaced 36mm apart and an FDA903D 45W audio amplifier. An 8Ω speaker that supports both local and cloud-based voice user interfaces.

This Amazon standard-compliant solution allows for rapid integration of the Alexa Voice service into embedded devices.

The main algorithmic flowchart is shown below and detailed information can be found in the attached file (Hybrid Quantization for Voice Command Recognition on Microcontrollers)

X-CUBE-LocalVUI implements a local speech recognition user interface based on audio capture and speech recognition. It integrates Sensory TrulyHandsfree™ (THF) and Sensory TrulyNatural™ (TNL) software.

The audio capture is based on STM32 peripherals and middleware. It shows how to capture audio from an on-board microphone via SAI.

The example application comes with preset speech recognition models and the user can easily update them with specific models. For the example, specific models can be defined using the Sensory Speech Center web tool.

Portability to a number of other STM32 microcontrollers and boards is possible.

►Scenario Application Diagram

►Photo of display board

►Program Block Diagram

►

Core Technology Advantage

High-performance STM32H7 microcontroller

With the performance and flash capacity of an ARM Cortex-M7 core, this highly integrated MCU manages high-end cloud-based voice UI features including fast wake word detection, advanced audio front end (AFE), and full connectivity stacking on cost-effective LQFP 100-pin packages with no additional peripherals or memory requirements.

Amazon Fully Qualified Software Reference Design

Fully qualified software for Amazon AVS for AWS IoT is fully functional and free of charge, with the exception of the evaluation version of the Alexa wake word component, which requires an Amazon license in the final product.

High Quality MP23DB01HP Microelectromechanical System Microphone

The ultra-compact, low-power, omnidirectional digital MEMS microphone consists of a capacitive sensing element and an IC interface with stereo operation. The element has a very high AOP in performance mode, a sensitivity range of ±1 dB, and a high SNR in all operating modes

Powerful FDA903D Audio Amplifier

This highly efficient 2 W single-bridge Class D amplifier with I45S inputs includes a high-performance D/A converter with high-performance output MOSFETs.

Maximize integration and cost-effectiveness with a POS BOM of < $10 in volume

►Program Specifications

STM32H753VIT6E high-performance MCU with 2 MB embedded Flash, 1 Mb

embedded SRAM and cost-effective LQFP package

-2.4 GHz Wi-Fi subsystem and Murata 1DX module in bypass mode

Coupled to ISSI IS25LP016D 2 MBytes NOR Flash Memory

-3 MP23DB01HP MEMS microphones with 36 and 30 mm spacing

-FDA903D Class D digital input automotive audio amplifier

-8 Ohm speaker

-4 RGB LEDs and 4 simple LEDs

-Joystick, reset and user buttons

-Highly modular mother/daughter board

-36x65 mm² small footprint, simple and cost-effective PCB design

-Amazon certified acoustic far-field and noisy environment support

-Local wake word detection

-Audio output and wireless upgrade