Michael Kagan, Mellanox
By exploiting deep learning on a variety of cost-effective platforms such as NVIDIA's Jetson Xavier AGX, DeepSig's OmniSIG sensor provides visibility, analytics, and actionable threat detection capabilities at the edge of wireless networks. Leveraging a scalable inference architecture, OmniSIG enables continuous monitoring of wide spans of the radio spectrum, numerous types of emitters, and fine-time resolution at a fraction of the engineering and deployment costs of traditional solutions. OmniSIG has also been integrated with several industry leading analysis tools such as 3dB Labs, Virtualitics, and Kibana to provide user-friendly analytical and dashboard tools for viewing and understanding wireless activity at the network edge.
Beyond detection and classification, DeepSig will also explore how our OmniPHY solution provides a new, radically different approach to communications, where baseband modem systems are learned and adapted to wireless channel measurements and use real-world hardware & channel effects.By providing a new class of wireless systems that is able to exploit more information in the wireless environment, OmniPHY systems can provide improved efficiency, resilience, power efficiency and adaptivity in complex, high-density, non-linear, hostile, and/or unique communications environments where previous systems were too brittle or inflexible for real-world conditions. Finally, we will discuss how Cloud RAN solutions can be significantly enhanced using learning in the BBU processing in order to improve performance by leveraging machine learning to replace key existing signal processing algorithms.