Nordic Thingy:53 accelerates IoT product prototyping
Nordic Semiconductor’s Thingy range now includes the Thingy:53, which is a multi-sensor prototyping platform with multi-protocol short range wireless connectivity and support for embedded machine learning. It is designed to shorten the development cycle for wireless proofs-of-concept and for prototypes with ML capabilities.
The Thing:53 combines Nordic Semiconductor’s nRF5340 dual-core Arm Cortex M-33 multi-protocol SoC, its nPM1100 power management IC power management IC, the nRF21540 front end module (FEM), a power amplifier / low noise amplifier (PA/LNA) range extender and multiple sensors, with embedded machine learning (ML) firmware.
The prototyping platform is equipped with a rechargeable 1350mAh Li-poly battery and multiple motion and environmental sensors. It supports Bluetooth Low Energy (Bluetooth LE), Thread, Matter, Zigbee, IEEE 802.15.4, NFC and Bluetooth mesh RF protocols, and comes with pre-installed firmware for embedded ML directly on the Thingy:53.
The nRF5340 SoC features a dedicated application processor and a network processor. The Arm Cortex-M33 application processor is clocked at 128MHz and ensures the prototyping platform can handle advanced algorithms and the complex computational tasks associated with embedded ML, explained Nordic. There is 1Mbyte flash and 512kbyte RAM.
The network processor is designed for power-efficient operation, said Nordic and features its own computational resources to ensure robust wireless connectivity without compromising application software execution.
Close collaboration between Nordic and tinyML specialist, Edge Impulse, has resulted in the Thingy:53 being supplied with Edge Impulse firmware and the nRF Edge Impulse mobile app. The firmware collects training and test data from the sensors on the Thingy:53 and forwards the data over the air using Bluetooth LE to a mobile app. The mobile app forwards the data to Edge Impulse Studio (a cloud-based development platform for ML) where it is used to build and test an embedded ML model. The ML model can then be deployed to the Thingy:53 over the air using the same mobile app. This app will also show inferencing results. The firmware and the app enable developers to rapidly collect data and test ML models without connecting any cables to the Thingy:53, said Nordic.
The nRF Programmer app supplied with the Thingy:53 simplifies prototyping by allowing the developer to select from pre-made firmware and then update the Thingy:53’s firmware directly over the air from an iOS or Android device. This capability means new firmware can be deployed at any time and anywhere, without needing a desktop computer.
Thingy:53 is also compatible with Nordic’s mobile apps, nRF Mesh, nRF Connect, nRF Toolbox, and nRF Blinky.
The Thingy:53 includes a six-axis inertial measurement unit and an additional low power accelerometer. The environmental sensors include temperature, humidity, air-quality and -pressure. The product also incorporates colour and light sensors, a buzzer and a pulse density modulation (PDM) microphone.
The embedded ML functionality enables developers to use the Thingy:53’s sensors in applications such as voice recognition or movement pattern detection. The accelerometer and the PDM microphone can also be used to wake the nRF5340 SoC from sleep when triggered by motion or sound events. This is useful for creating low power embedded ML applications because Thingy:53 can remain in sleep mode to extend battery life when there are no external stimuli.
Thingy:53 also has a USB-C charge-and-data port, and a Qwiic, Stemma and Grove-compatible four-pin JST connector. Every Thingy:53 is delivered with a debug and current measurement board, which provides access to relevant pins that would otherwise be unavailable, for troubleshooting application code and monitoring power consumption with additional debug tools.
The Thingy:53 is available now from Nordic’s distribution partners.