Shenzhen-based Wesion Technology’s Khadas Project has recently launched an entire ecosystem of RK3399-based development SBCs and add-on products on Indiegogo, which are expected to ship by January of next year. Khadas describes their Edge line as “a next-generation single-board computer, that was made with the dream of helping makers tackle the latest AI and ML problems.” Actually, one of the boards was explicitly designed for AI and machine learning development, although all three have the capacity to handle those applications.
The Khadas Edge is the company’s initial offering and is outfitted with Rockchip’s RK3399 SoC (dual-core Cortex-A72/quad-core Cortex-A53-Mali-T860), along with 2GB of LPDDR4 RAM, and 16GB of eMMC 5.1 storage. It also features HDMI 2.0a, Gigabit Ethernet, 802.11ac Wi-Fi, Bluetooth 4.1, a USB 3.0 port, a USB 2.0 port, 2X USB Type-C ports, and an SD card slot.
The Edge sports a 314-pin edge connector as well, which enables it to attach to the company’s Captain Carrier board for additional functionality. This version comes in three flavors — Basic (outlined earlier), Pro (adds 4Gb of RAM/32Gb of eMMC), and Max (4Gb of RAM/128GB of eMMC), with both higher-ended versions adding Bluetooth 5.0.
The Edge-V has all of the same feature sets as the Edge but offers increased I/O options, including 40-pin GPIO (removing the gold fingers connection), USB 3.0, USB 2.0, USB-C (3.0/DP), and USB-C (PD). It also sports MIPI-RX/TX, MIPI-TX, MIPI-RX, HDMI 2.0, M.2 socket (PCI-E 2.0), E-DP display connector, TF card slot, and a pair of I-PEX Wi-Fi/Bluetooth connectors. The Edge-V even packs a couple of sensors (IR, gesture), RTC battery header, and a touch panel port for interactive displays.
According to Khadas, the Edge-1S is still under “heavy” development and have yet to release the schematics for the board; however it sports the same feature set as the other two, only it comes equipped with an RK3399 Pro SoC with an embedded 3.0 TOPS Neural Network Processor (with 8/16-bit processing). This SBC will support TensorFlow, Caffe, MLlib (among others), and will be able to handle AI/machine learning applications, plus AR, facial recognition, and object tracking, among a host of others wicked features.
Khadas’ Captain Carrier Board is designed to host the Edge and Edge-1S, giving the SBCs increased feature sets, such as a micro SD slot, an M.2 2280 slot with support for NVMe solid-state storage, an accelerometer, gyroscope, and APDS-9960 sensor for gesture detection. It’s outfitted with a 40-pin GPIO header, 3.5mm audio jack, dual game-pad buttons, and more.
Aside from all that, Khadas is offering a host of add-on options for the Edge line, including Li-Po battery, heatsink, 5-inch LCD touchscreen, and a 13-megapixel IMX214 camera module, which are available on the company’s Indiegogo campaign page. Pledge prices are wide-ranging and start with $99 for the Edge Basic, and only increase from there but all seem well worth the money.