DeepSig’s OmniPHY® 5G software replaces standard algorithms within the 5G-NR high-PHY, which can be readily leveraged in OpenRAN and vRAN solutions in order to increase performance through improved link margin and reduced power consumption. By leveraging data within the uplink, we are able to drop in very low latency neural networks to perform estimation and equalization to improve link performance. Below we illustrate where this module drops into a typical vRAN stack.
This low latency HI-PHY enhancement can also be coupled with a RIC xAPP to monitor and continuously improve algorithm performance over time. We provide a unique set of RIC E2 enhancements which allow xApp’s to extend their performance optimization even deeper into the communications stack.
The DeepSig 5G-AI Lab is operating commercial UE calls over the air with OmniPHY® 5G today, streaming content over AI-Native drop in replacement software.
Massive MIMO is an exciting core technology within 5G, which allows for dense spectral re-use through improved beam management and spatial multiplexing of multiple users in a wide range of deployment scenarios. While this exciting capability provides significant capacity advantages, it is notoriously power hungry and sub-optimal in many of the early algorithms and deployment simplifications. OmniPHY® 5G provides a set of software algorithm enhancements which leverages data in order to improve mMIMO system performance, and to drive down power consumption, making deploying and operating 5G Massive MIMO systems affordable, efficient, and autonomous.
OmniSIG® can be further deployed within the context of the 5G vRAN in order to help optimize for resource allocation, provide resilience and avoidance of RFI, and to help private networks and emergency 5G networks operate and share spectrum smarter, more reliability and more autonomously. OmniSIG® deployments in the RIC as a set of edge-sensing and xApp’s provides the intelligent spectrum aware future of the RAN in today's systems and provides a powerful foundation for future RAN intelligent automation.