text.skipToContent text.skipToNavigation

 

NXP

NXP EdgeReady SLN-ALEXA-IOT Development Kit

MCU-Based Solution for Alexa Integration

NXP’s EdgeReady SLN-ALEXA-IOT is an MCU-based solution for Amazon Alexa Voice Service (AVS) Integration for AWS IoT Core. The kit leverages the NXP i.MX RT106A audio crossover processor, enabling developers to quickly and easily build Alexa into their products.

The development kit (SLN-ALEXA-IOT) comes with Amazon qualified software for an out-of-the-box Alexa experience with easy Wi-Fi® and Alexa account provisioning using Frustration Free Setup.

NXP’s MCU-based Alexa IoT solution provides device makers with a turnkey hardware and software reference design based on NXP’s i.MX RT106A MCU. The development kit features a machine learning far field audio front end that includes algorithms for deep learning based filtering, beamforming, echo cancellation and barge-in capabilities, enabling use in acoustically challenging environments with only two (optionally three) microphones.

Also included is Amazon’s Alexa wake word engine. The out-of-the-box the solution development kit, SLN-ALEXA-IOT, is easily connected to the user’s Wi-Fi network and Alexa account using Amazon’s Frustration Free Setup (FFS). The solution has been AVS and FFS qualified by Amazon and is built on Amazon Common Software (ACS) middleware and FreeRTOS for easy integration with the Alexa Voice Service and AWS IoT, with future proofing for new Amazon ACS based applications.

OEMs can efficiently add Alexa or other cloud based voice assistants to their smart home and smart appliance products, often without requiring the services of an Alexa System Integrator.

The SLN-ALEXA-IOT includes:

  • Machine learning audio front end for far field voice
    • Deep learning based filtering and beamforming
    • Acoustic echo cancellation (barge-in)
  • Wake word inference engine
  • Playback processing
  • Codecs, media player/streamer
  • MQTT, lwIP, TLS
  • Running on FreeRTOS and AWS IoT Core
  • Built on Amazon Common Software (ACS) middleware
  • Amazon Frustration Free Setup (FFS) discovery and onboarding
  • Alexa locales
  • Alexa smart home
  • All hardware drivers, including Wi-Fi and Bluetooth
  • Supported by MCUXpresso SDK, IDE and configuration tools

Target Applications

Any device that can add Alexa to anywhere in the home, to provide ubiquitous access to Alexa without requiring Echo speakers everywhere, on a countertops and walls.

  • Smart home:
    • Smart switches, lighting/shade/fan controls and smart plugs/outlets
    • Thermostats
    • Alarm panels and glass break sensors
    • Smoke and CO detectors
    • Set top boxes, TVs and audio video receivers (AVRs)
    • Home gateways, routers access points and mesh Wi-Fi systems
    • Garage door openers and access panels
    • Room air conditioners and (de-) humidifiers
    • Low cost smart speakers
  • Smart appliances
    • Whitegoods (fridge, oven, washer, dryer, cooktop, vent hood, etc.)
    • Countertop (microwave, coffee maker, food processor, multicooker, etc.)

Advantages of the NXP i.MX RT106A for Alexa

  • Lowest cost
    • Arm Cortex-M7 @600 MHz + 1 MB SRAM – replaces host MCU
    • No external DSP or wake word engine with, integrated codec
    • Using MCU halves the cost of a Linux based full AVS implementation
      • Eliminates SDRAM, eMMC Flash, PMIC and uses 4 layer board
      • Cloud offload reduces frequency, size and cost of OTA updates
  • Fastest & easiest – concept to production in as little as four months
    • Familiar MCU+RTOS platform (no Linux learning curve)
    • Turnkey solution – one stop shop – includes all software
      • No voice or audio expertise necessary
        • Machine learning far field audio front end (AFE)
      • No System Integrator needed, no third-party DSP or AFE devices
      • Plug & play out-of-box-experience, like an Amazon Echo speaker
      • AVS qualified, meets/exceeds performance of Amazon’s Echo Dot
    • World-wide availability and support
  • Most flexible
    • Only solution supporting 360° far-field with option for dual band Wi-Fi
  • New in SDK release 2.8.1
    • Frustration Free Setup
    • Alexa locales (for world-wide usage)
    • Alexa Smart Home API
    • SDK for BOM cost optimized reference design

Alexa Built-in Solution Reference Design Block Diagram

NXP – Alexa Built-in Solution Reference Design Block Diagram

The reference design for this solution is a BOM cost optimized version of the SLN-ALEXA-IOT development kit.

At the heart of the solution is the i.MX RT106A MCU, operating at speeds of up to 600 MHz. The reference design includes microphones and an audio amplifier to drive a speaker. Wi-Fi and Bluetooth/BLE drivers have been pre-integrated for connectivity.

  • The reference design uses 16MB of QSPI Flash with the i.MX RT106A
    • The i.MX RT106A can be used to replace the host MCU in an existing product design that does not have voice
  • Two digital PDM microphones are required for 180° far-field designs
    • For on wall applications, or on the front panel of a large appliance, etc.
  • Optionally a third microphone can be added for 360° far-field voice acoustics
    • Where the device might be placed in the middle of a room
    • _Using three microphones requires the use of HyperFlash
  • A speaker & analog amplifier are required for audio feedback, driven by the on-chip digital to analog converter (DAC), the MQS module of the i.MX RT106A
  • A Wi-Fi + BLE module is included

Alexa Built-in Solution – Software Block Diagram

NXP – Alexa Built-in Solution Software Block Diagram

NXP’s EdgeReady solution for Alexa Built-in includes production-ready software for customers wanting to add Alexa to their products. The software is optimized to run on an MCU implementation, and leaves headroom for the CPU to run other control functions. Running on FreeRTOS, this solution includes Amazon’s Alexa wake word engine and a machine learning audio front end (AFE) that provides echo cancellation, noise suppression, barge-in and beamforming to ensure audio accuracy in noisy environments.

  • All software provided as source code except
    • Machine learning audio front end
    • Alexa wake word engine
  • No third-party engagements required
    • Turnkey, FCC certified and Amazon qualified hardware and software.