The AIR8201-B from Deepwave Digital is an Artificial Intelligence Radio Transceiver (AIR-T) that operates from 300 MHz to 6 GHz (Tx) and 30 MHz to 6 GHz (Rx). It has been specifically designed as a complete deep learning/RF sensor for operation as an edge-compute device in harsh environments. The transceiver integrates an expanded receiver frequency range (down to VHF), 256-core NVIDIA Jetson Tx2i processor with GPU for parallelized processing, Xilinx Artix 7 200T FPGA, a GNSS disciplined oscillator for timing, expanded SSD data storage (512 GB), and a CPU for control, I/O, and software applications. The various processors integrated into this transceiver can be used to perform all deep learning inferences on the radio, without the need to reach back to a data center.
The AIR8201-B is versatile and can function as a highly parallel SDR, data recorder, or inference engine for deep learning algorithms, thereby allowing deep learning and GPU accelerated application to occur on the radio, at the edge. It incorporates a 2x2 MIMO RFIC chip that provides a gain of 52 dB and has a noise figure of 2.2 dB in the Rx chain. The GPSDO has a timing accuracy of 40 ns with a holdover stability of less than ±400 ns at 1 hour. It provides up to 8 anti-jam countermeasures against CW signals and are designed to receive signals in the GPS, GLONASS, Galileo, QZSS, and SBAS bands. This transceiver supports a data speed of up to 1 Gbps over Ethernet cable and offers other I/O connectivity options such a USB 3.0, USB 2.0, and GPS sync.
This AI Radio Transceiver is available in an environmentally rugged enclosure that measures 12.2 x 11.0 x 2.4 inches and is ideal for automatic speech recognition, signal classification, edge data processing, signal intelligence, spectrum administration, and electronic warfare applications.