BeagleBone AI puts devs in the machine learning fast lane
Revealed at the Embedded World expo in Nuremberg, Germany, last week, the BeagleBone AI has been developed to fill the gap between small single board computers and bigger, more powerful industrial computers.
At this stage, the non-profit BeagleBoard Foundation – which is perhaps best known for its more capable alternatives to the Raspberry Pi – isn't giving too much away. The AI board is a kind of smaller X-15, but with a few sacrifices made to keep the price down.
The BeagleBone AI is built around a Texas Instruments AM5729, supported by 1 GB of RAM – that's half the system memory of the X-15 – and 16 GB of onboard eMMC flash storage.
This is a board created for artificial intelligence projects, so as well as treating the new board to four programmable real-time units, the Foundation has also included four embedded vision-engines, a TI C66x digital signal processor, and has pre-installed machine learning tools.
This model boasts two expansion headers, USB-C for power and data, Gigabit Ethernet and cooked-in "high speed" Wi-Fi too.
The Foundation is hoping to make the BeagleBone AI available by April for around US$100.
Source: BeagleBoard Foundation