MaxRefDes178 # “is a cube camera reference design based on the MAX78000 and MAX32666 microcontrollers to help AI in edge device designers accelerate their proof-of-concept to market,” the company said.
Measuring 41 x 44 x 39mm, it has a pair of PCBs inside (CAD files are provided), one for the AI and one for connectivity.
The AI board has two MAX78000 ICs – Arm Cortex M4F MCUs with Convolutional Neural Network Acceleration – one is connected to the image sensor for image and video applications, the other with an integrated digital microphone and a stereo audio codec for speech and audio. “This configuration enables the MaxRefDes178 to run two different AI models at the same time: for example, parallel to face recognition and keyword recognition,” said Maxim.
The connectivity board implements wired and wireless communication as well as user interaction via a capacitive touch LCD screen on the back of the cube. This board also contains: ceramic antenna, accelerometer, gyroscope, battery charger, audio input, audio output, micro SD card slot and on-board flash memory – the latter for storing pictures, AI networks and audio clips.
Control and wireless communication on the connectivity board is via the Bluetooth-enabled dual Cortex-M4-Core-MAX326666-MCU – with one core managing the Bluetooth Low Energy and USB software stacks, while the other controls the hardware and data between the TFT -LCD and memory transmits, audio codec, USB and BLE radio.
The device is pre-programmed with an application and can simply be connected to the power supply via a USB Type-C cable to start it up.
An associated Android app is available from the Google Play Store.
An external circuit board and cables (all included in the kit) are required for programming and debugging. It uses the MAXDAP Type-C DAPLink-based debugger. OpenOCD is installed within the development tool chain.
“Machine learning promises a lot: that machines understand what they see and hear, how humans can understand and make more autonomous decisions,” said Kris Ardis, director of Maxim, and claims: “Until the MAX78000, the embedded world was left behind” because you got AI on Edge could not implement with limited performance, cost and size. “