Ultra low power AI sensor board for edge machine learning

6th July 2020

Eta Compute

Alex Lynn

0 0

Eta Compute has announced the debut of an ultra low power AI Sensor Board, the ECM3532 AI Sensor Board. Designed specifically for machine learning at the edge, the board reduces the initial phase of smart sensor development using its TENSAI SoC, including feasibility, proof of concept, and board design, from several months to weeks or even days.

The company also announced it is hosting a free online workshop with Edge Impulse July 14th, 2020. The first 50 qualified developers planning or working on projects will receive a complementary ECM3532 AI Sensor Board. Developers should apply on the Eta Compute website. The board can be purchased through Digikey.

“Our fully integrated ECM3532 AI Sensor Board reduces risk, complexity, and costs for developers creating low power smart sensors and accelerates their time to proof of concept,” said Ted Tewksbury, CEO of Eta Compute. “For the first time, they can rely on a single, fully integrated board complemented by Edge Impulse’s comprehensive SaaS development platform to drive innovative IoT solutions to market.”

“Designing low-power devices with embedded machine learning has been a huge challenge,” added Zach Shelby, CEO of Edge Impulse. “The new Eta Compute AI Sensor Board opens up endless application possibilities for developers without compromising on power consumption, and with full support on Edge Impulse we reduce development time to days instead of months.”

Previously, smart sensor design required developers to kluge multiple power-hungry boards together. With Eta Compute’s low power ECM3532 AI Sensor Board, virtually every function needed, including data acquisition, is integrated into a single board that functions for months on a single battery cell while performing machine learning algorithms.

It embeds two PDM microphones, one pressure and temperature sensor, and one 6-axis MEMS accelerometer and gyroscope. Its small form factor (1.4x1.4”), the embedded battery socket and Bluetooth connectivity with the A31R118 from ABOV Semiconductor, a leader in Bluetooth Low Energy solutions, make it an ideal asset for IoT deployment and field testing application prototypes. When needed, the board’s expansion connector simplifies the addition of other RF interfaces.

Eta Compute’s ECM3532 AI Sensor Board is optimal for implementing machine learning applications in IoT and low power edge devices including:

Voice activation and voice commands in consumer, home, and building automation products
Asset tracking and monitoring and predictive maintenance in industrial products
Context awareness in consumer and industrial applications