Artificial Intelligence Radio–Transceiver review

By Marc de Vinck. Posted about 2 years ago.

With three integrated processors, the Artificial Intelligence Radio-Transceiver (AIR-T) is capable of harnessing all that power, and leverage machine learning, to allow for fully autonomous signal identification and interference mitigation. There is even an embedded GPU for real-time, wide-band digital signal processing and deep learning.

And yes, this is all getting way above my skill set, but that also makes it incredibly intriguing – to learn more about the state-of-the-art in machine learning and SDR technologies. The AIR-T comes preloaded with AirStack, which includes all the components necessary to run an Ubuntu-based OS, the AIR-T-specific drivers, and the Xilinx Artix-7 FPGA firmware. If you’d like to integrate TensorFlow, Caffe, Keras, or PyTorch with an SDR, along with six CPU cores, then you should check out the AIR-T. Just make sure that you really need its power, because it comes at a fairly high price.