DeepSig, Inc.
DeepSig develops AI and machine-learning software and research for wireless communications. Primary focus areas include neural-network implementations of radio physical-layer functions for virtualized RAN, machine-learning spectrum sensing and RF classification, open-source CU/DU software for Open RAN, and AI-enhanced massive MIMO techniques. The company publishes historical RF datasets and model repositories for research under a Creative Commons non-commercial share-alike license.
Industries
N/A
Nr. of Employees
small (1-50)
DeepSig, Inc.
Patents
Radio frequency band segmentation, signal detection and labelling using machine learning
US-12373715-B2
View Details
Radio frequency band segmentation, signal detection and labelling using machine learning
US-12373715-B2
View DetailsProducts
Neural receiver software for 5G vRAN
Software implementing a neural-network upper PHY for virtualized RAN deployments to improve channel estimation, equalization and uplink throughput compared with conventional PHY processing.
ML-driven spectrum sensing software
Real-time wideband spectrum sensing and analytics software for detection, classification and localization of RF emissions for operational and security use cases.
Model training and deployment studio for spectrum ML
Software environment for curating RF datasets, training models, and deploying spectrum-sensing ML models to edge and cloud targets.
Neural receiver software for 5G vRAN
Software implementing a neural-network upper PHY for virtualized RAN deployments to improve channel estimation, equalization and uplink throughput compared with conventional PHY processing.
ML-driven spectrum sensing software
Real-time wideband spectrum sensing and analytics software for detection, classification and localization of RF emissions for operational and security use cases.
Model training and deployment studio for spectrum ML
Software environment for curating RF datasets, training models, and deploying spectrum-sensing ML models to edge and cloud targets.
Services
Integration of neural-network based upper-PHY software into virtualized RAN components to improve channel estimation, equalization and uplink throughput.
Continuous ML-powered sensing and analytics to detect, classify and localize emitters for network optimization, interference mitigation and security monitoring.
Tooling and workflows for training, validating and deploying ML models for spectrum sensing and RF applications to edge and cloud targets.
Integration of neural-network based upper-PHY software into virtualized RAN components to improve channel estimation, equalization and uplink throughput.
Continuous ML-powered sensing and analytics to detect, classify and localize emitters for network optimization, interference mitigation and security monitoring.
Tooling and workflows for training, validating and deploying ML models for spectrum sensing and RF applications to edge and cloud targets.
Expertise Areas
- AI-native physical-layer communications
- Machine-learning spectrum awareness and RF classification
- Open RAN CU/DU software development and integration
- Massive MIMO and ML beamforming optimization
Key Technologies
- Deep neural networks for PHY
- ML-driven spectrum sensing
- Neural receiver algorithms for vRAN
- Massive MIMO beamforming algorithms
News & Updates
Announcement of a DoD-funded program led by InterDigital with DeepSig as an industry partner to develop and validate AI-enabled spectrum coexistence technologies for civil and military applications.
Over-the-air demonstration of a learned, pilotless air interface and digital twin simulations showing spectral and power efficiency gains; demonstration used GPU-accelerated platforms for L1 processing.
Awarded NTIA grant to accelerate AI-enhanced massive MIMO development in partnership with industry chipset and radio unit vendors to optimize beamforming for Open RAN.
Exhibited AI-native modem and spectrum awareness demonstrations at NVIDIA GTC Washington, D.C., showcasing learned PHY and small-form-factor spectrum sensing on GPU-accelerated platforms.
Chosen with partner organizations to lead an open-source, carrier-grade CU/DU initiative to accelerate 5G/6G open RAN software development.
Awarded a grant by the National Telecommunications and Information Administration to develop AI-enhanced massive MIMO solutions for radio units in Open RAN.
Announcement of a DoD-funded program led by InterDigital with DeepSig as an industry partner to develop and validate AI-enabled spectrum coexistence technologies for civil and military applications.
Over-the-air demonstration of a learned, pilotless air interface and digital twin simulations showing spectral and power efficiency gains; demonstration used GPU-accelerated platforms for L1 processing.
Awarded NTIA grant to accelerate AI-enhanced massive MIMO development in partnership with industry chipset and radio unit vendors to optimize beamforming for Open RAN.
Exhibited AI-native modem and spectrum awareness demonstrations at NVIDIA GTC Washington, D.C., showcasing learned PHY and small-form-factor spectrum sensing on GPU-accelerated platforms.
Chosen with partner organizations to lead an open-source, carrier-grade CU/DU initiative to accelerate 5G/6G open RAN software development.
Awarded a grant by the National Telecommunications and Information Administration to develop AI-enhanced massive MIMO solutions for radio units in Open RAN.