SLN-SVUI-IOT Turnkey solution introduction

[#1]

SLN-SVUI-IOT Turnkey solution introduction

1. Abstract

NXP SLN-SVUI-IOT EdgeReady solution for both local and online voice control leverages the i.MX RT106V crossover MCU with integrated Voice Intelligent Technology (VIT) offering a voice user interface for touchless applications.

This ultra-small form-factor, production-ready hardware design comes with fully integrated software running on FreeRTOS for quick out of-the-box evaluation and proof of concept development. This turnkey solution minimizes time to market, risk, and development effort enabling OEMs to easily add voice to their Industrial and IoT products.

Fig 1

2 Key Feature

Low cost

• Arm Cortex-M7 – 600 MHz + 1 MB SRAM
• No external DSP or wake word engine, integrated codec
• Replaces host MCU (not an add on solution)
• Less than half the cost of a Linux based implementation on an MPU
  • Eliminates SDRAM, eMMC Flash, PMIC and uses 4-layer board

Fastest & easiest – concept to production in less than 6 months

• Familiar MCU+RTOS platform (no Linux learning curve)
• Turnkey solution – one stop shop – includes all software
  • No System Integrator needed, no third-party engagements
  • No voice or audio expertise necessary – machine learning far field AFE
  • Includes proven phrase spotting automatic speech recognition (ASR) engine
  • Plug & play, out-of-box-experience
  • Similar far field voice performance to Amazon’s Echo Dot
• 2 or 3 mic. support for 180° or 360° far-field implementations
• Worldwide availability & support

3. Local Voice Control Target Applications

Anywhere that needs hands-free, private voice control without cloud connectivity

Smart Home

• Smart lighting, shade and fan controls
• Smart switches, dimmers, plugs and outlets
• Thermostats, room air conditioners and (de-)humidifiers
• Alarm panels, glass break sensors, smoke & CO detectors
• Set top boxes, home gateways and routers
• Garage door openers and access panels
• Smart toys

Smart appliances

• Major (fridge, oven, washer, dryer, cooktop, vent hood, wine cooler etc.)
• Countertop (microwave, coffee maker, food processor, multi cooker etc.)

Smart buildings and industrial

• Elevators
• Intercom systems in multiple dwelling units
• Vending machines
• Industrial automation and hands-free process control

Fig 2

4. Production Grade, Certified and Qualified reference systems

Fig 3

5. Hardware and software situation

SLN-SVUI-IOT hardware highlights:

• Up to 600 MHz (528 MHz default) Cortex-M7 MCU core
• 1 MB of on-chip RAM (512 kB TCM)
• Multiple microphone topologies:

         – Two PDM mics on main board (not active by default)

         – Two PDM mics on extension board (not active by default)

         – Three I2S mics on extension board (active by default)

• 3 W mono filter-less class-D amplifier Wi-Fi/Bluetooth combo chip (intended to be used for OTA updates, if needed by customers)
• Integrated speaker
• GPIO expansion headers

Fig 4

Fig 5

SLN-SVUI-IOT software highlights:

• Two-stage bootstrap and bootloader allowing flexibility in customer’s implementation
• Secure boot flow with high assurance booting (HAB)
• Over-the-wire (OTW) update via UART
• Automated manufacturing/reprogramming tools
• Speech recognition engine by deep learning
• Audio front end (AFE) for far-field automatic speech recognition (ASR)

Fig 6

The SLN-SVUI-IOT kit is supported by a comprehensive and free-of-charge enablement suite from NXP and its partners including:

• MCUXpresso development tools
• Hardware design files
• Local voice application software source code
• Software audio tuning tools
• Documentation
• Training material

6. Smart Voice UI technology

Part numbers for smart voice UI:

RT1062: Voice Seeker (no AEC) + VIT

RT106V: Voice Seeker (with AEC) + VIT

RT106C: Voice Seeker (with AEC) + Cyberon DSMT

6.1 Voice Seeker

Multi-microphone audio front-end signal processing solution for low-power, always-on devices. It provides multi-mic beamforming, noise suppression, and multi-channel ecoustic echo cancellation, enabling high performance far-field speech pickup.

• nxp.com/VoiceSeeker
• VoiceSeeker Overview Video

Key Features/Benefits

• Flexible microphone geometries are supported
• Beamforming, Noise Reduction, Dereverberation, Payload capture
• Direction of Arrival Indication accurate up to 1 degree
• Optional Multi-channel Acoustic Echo Cancellation available
• Integrates easily with VoiceSpot and VIT Engines 
• Standard Enablement without AEC included in MCUXpresso SDK

Fig 7

6.2 Voice Intellgent Technology

Voice Intelligent Technology (VIT) Wake Word and Voice Command Engines provide free, ready to use voice UI enablement for developers. It enables customer-defined wake words and commands using free online tools. The library and voice control software package is delivered via the MCUXpresso SDK or Linux BSP.

Based on deep learning speech recognition technologies, this software package provides a complete wake word and voice command solution. VIT can be easily configured with VoiceSeeker, a multi-mic audio front end supporting far-field operation. The VIT Wake Word and Voice Command Engines are available royalty-free on several platforms including Arm^® Cortex^®-M7, M33, A-53 or Cadence Xtensa^® HiFi 4 and Fusion F1 cores

https://www.nxp.com/vit

Fig 8

Feature:

• VIT is based on state-of-the-art Deep learning and speech recognition technologies
• VIT is a complete NXP IP for Voice enablement on any relevant NXP platform and is free for customer use (binary library provided)
• Wake word model creation with Text to Model (no audio database required)
• Custom commands using Text to Model
• Large vocabulary available for Text to Model
• English, Chinese (Mandarin), French, German, Italian, Japanese, Korean, Spanish, Turkish, language support: in production on vit.nxp.com
• Up to 3 Wake Words supported in parallel
• Current limit of 30 commands per model

Fig 9

6.3 Cyberon DSMT

DSpotter modeling tool (DSMT) is a user-friendly tool for creating customized models using customer-defined wake words and commands.

Note: This tool requires an Internet connection.

Fig 10

Fig 11

To create a model, follow the steps below:

Log in with your credentials. For access, contact local‑commands@nxp.com. Ensure to specify the following details in the email:

• Name
• Email ID
• Company name
• MAC address

7.Smart Voice UI solution advantages Summary

Fig 12