AAEON's BOXER-8224AI Is a Slimline NVIDIA Jetson Nano Edge AI Board for Drones and Robotics
Designed to bring on-device machine learning to more compact projects, this Jetson Nano carrier brings features out to wafer connectors.
Industrial and embedded computing specialist AAEON Technology has launched a new device aimed at drone and robotics projects that need some high-performance but low-power on-device artificial intelligence, powered by the NVIDIA Jetson Nano system-on-module (SOM): the BOXER-8224AI.
"Ushering in a more flexible way to bring applications to market, the BOXER-8224AI is an edge AI solution with a difference," AAEON claims of its design. With the board’s coastline I/O [Input/Output] replaced with pin-wafers, including MIPI-CSI camera connectors, it maintains the functionality of a system-level solution with the flexibility of a board, making it the smart choice for applications that require sophisticated interfaces in limited installation spaces, such as drones."
The BOXER-8224AI, brought to our attention by CNX Software, is not technically a single-board computer — but certainly fills the same niche. It's a carrier board which comes bundled with NVIDIA's Jetson Nano system-on-module, the same model as found in the original Jetson Nano Developer Kit — giving it a quad-core 1.43GHz Arm Cortex-A57 processor, 128-core Maxwell graphics processing unit (GPU) with CUDA support, and 4GB of RAM.
Broken out from the SOM's feature set is a mini-HDMI 2.0 port for video connectivity and a seres of low-profile wafer connectors offering four USB 3.2 Gen. 1 ports, two gigabit LAN connections, two RS232/485 ports, UART, two I2C buses, six general-purpose input/output (GPIO) pins, two MIPI Camera Serial Interface (CSI) ports, a 5V power output, and a 12-24V power input. For storage, the board bundle comes complete with a 16GB eMMC module with microSD expansion, and there's an M2 E-key slot for an optional Wi-Fi/Bluetooth module.
It's the combined device's small size which will appeal for space-constrained robotics and drone projects: including the carrier board and SOM, the BOXER-8224AI measures just 4.72×3.1×0.86" and weighs 0.23lbs — helping to keep more weight free for payload. Given airflow, the company claims the bundle can run in temperatures between -4°F and 140°F — even when using the Jetson Nano's GPU to run on-device machine learning workloads like computer vision.
The bundle, which includes carrier board and SOM, is now available on the AAEON store for $525.