Reference Design:

BrickML

A standalone tool for measuring and monitoring machine health using machine learning at the edge.

Designed and built in conjunction with

A readily deployable plug-and-play edge ML design

BrickML is a low-power high-performance self-contained embedded device designed to run machine learning operations at the edge in industry settings.

Specifically built to be plug-and-play, BrickML is quickly implemented. Connect it to motors, pumps, or other machinery to measure machine performance and create immediate actionable insights.

Designed and built in conjunction with Reloc, BrickML module features on-board compute, sensing and communication capabilities, all packed in one compact industrial-grade dust and moisture resistant form factor. It is fully integrated into Edge Impulse's industry-leading edge ML platform, allowing for the easy addition of advanced AI algorithms.

BrickML module minimizes development time and accelerates speed to market by providing customers with both a highly integrated flexible kit for custom development as well as ready-to-use trained edge ML models.

Main applications

Embedded Machine  Learning

Inference at the Edge

Anomaly
Detection

Voice and Sound Classification

Motor Current Signal Analysis

Environmental Condition Analysis

Connected Intelligent Sensors

Specifications

Power Supply: 2.5V to 5.0V

Standby Current:1uA

Connectivity:Wired: USB Full-Speed, Ethernet, CAN
Wireless: BLE 5.1 (output 2.2dBm),
LTE (expansion)

Expansion: CAN, UART, I2C, SPI

Size40.0mm x 23.0mm x 5.0mm

Operating Temperature: -40°C to +85°C

Certifications:
IP67 Enclosure (waterproof/dust proof)
FCC & CE wireless certified
RoHS-compliant

Processing and Storage

32-bit ARM® Cortex®-M33 core with FPU

200 MHz clock speed

Up-to 2MB code flash memory, 8KB data flash and 512KB RAM

128Mbit on-board serial flash

SD card storage

Bluetooth 5.1 core specification compliant BLE transceiver and link layer

USB 2.0 Full-Speed device connector

Expansion pinout: Ethernet 10/100, USB 2.0 High-Speed, CAN, LTE, UART, I2C, SPI

Sensors

Audio:
Knowles SPH0641LU4H-1 microphone (100hZ – 80,000hZ frequency range)

Current: Voltage input to ADC for MCSA applications

Motion tracking:
Bosch BNO055 9-axis inertial sensor featuring 14-bit accelerometer, close-loop triaxial 16-bit gyroscope, and triaxial geomagnetic sensor

Humidity and temperature: Renesas HS3001 high-performance temperature/humidity sensor

Available from Mouser

For more info, get in touch with sales@edgeimpulse.com

Get started with Edge Impulse