VRAN NEXT — SENSING PLATFORM
5G Edge Connect
OranSense CUDA-Accelerated RAN brings together the OranSense software for 5G and AI frameworks and the NVIDIA accelerated computing platform, enabling TCO reduction and unlocking infrastructure monetization for telcos.
PLATFORM OVERVIEW
The Intelligent RAN Platform
The OranSense Sensing Platform is a fully AI-native, O-RAN compliant software stack that transforms commodity hardware into a high-performance, AI-augmented radio access network. Built on CUDA-accelerated compute, it delivers carrier-grade performance with the flexibility of open, disaggregated architecture.
ARCHITECTURE
A Layered Intelligence Stack
The platform is organized into three tightly integrated layers — each purpose-built to handle a distinct tier of the network intelligence hierarchy.
Sensing Layer
RF Signal Acquisition & Processing
Continuous wideband RF sensing across licensed and unlicensed spectrum. ISAC (Integrated Sensing and Communication) radar returns are fused with baseband I/Q data to build a real-time environmental model at each cell site.
Inference Layer
Edge AI & Model Execution
CUDA-accelerated inference engines run OranSense proprietary models directly on the DU/CU hardware. Sub-millisecond decision loops enable real-time beamforming optimization, interference cancellation, and anomaly detection without cloud round-trips.
Orchestration Layer
xApp & rApp Lifecycle Management
A fully O-RAN compliant near-RT RIC and non-RT RIC host OranSense xApps and third-party rApps. The orchestration layer handles policy enforcement, model versioning, A/B testing, and federated learning coordination across the RAN fleet.
KEY CAPABILITIES
What the Platform Delivers
CUDA-Accelerated Baseband
GPU-native L1/L2 processing delivers 3–5× the spectral efficiency of CPU-only stacks on identical hardware.
O-RAN Open Interfaces
Full E2, A1, and O1 interface compliance enables multi-vendor interoperability and third-party xApp integration.
Autonomous Self-Optimization
Closed-loop AI agents continuously tune power, tilt, and handover parameters — eliminating manual drive-test cycles.
InfraSense Fusion
Passive RF sensing data is fused with network KPIs to generate environmental intelligence layers for smart-city applications.
Federated Learning
Privacy-preserving model training across distributed RAN nodes — no raw data leaves the edge.
Digital Twin Integration
Live network state is mirrored into an NVIDIA Omniverse digital twin for simulation, what-if analysis, and predictive maintenance.
INFRASENSE — TELECOM AS A SENSOR NETWORK
Every 5G Tower Is Now a Sensor
InfraSense is OranSense's passive sensing layer that transforms existing 5G base stations into a city-scale environmental intelligence network — with zero additional hardware. By fusing RF signal patterns, ISAC radar returns, and edge AI inference, InfraSense delivers sub-second situational awareness across any geography already covered by 5G.
SENSING INTELLIGENCE
City-scale awareness from infrastructure you already own
Passive RF Sensing
Continuous wideband monitoring of licensed and unlicensed spectrum. InfraSense captures signal reflections, interference patterns, and propagation anomalies in real time — without transmitting a single additional waveform.
ISAC Radar Fusion
Integrated Sensing and Communication (ISAC) radar returns are fused with baseband I/Q data at the DU. The result is a high-resolution environmental model — object detection, crowd density, and motion tracking — derived entirely from the existing RAN.
Sub-Second Edge Inference
CUDA-accelerated AI models run directly on the cell-site GPU fabric. Decision loops close in under one millisecond — enabling real-time alerts, autonomous network adjustments, and live data feeds to smart-city platforms without cloud round-trips.
Multi-Layer Data Fusion
InfraSense correlates RF intelligence with network KPIs, weather telemetry, and third-party IoT streams. The fused output is published as structured intelligence layers consumable by city dashboards, emergency services, and enterprise applications via open APIs.
Zero Additional Hardware
InfraSense activates on existing 5G infrastructure through a software-only xApp deployment. MNOs and neutral hosts unlock a new sensing revenue stream from assets they already own — no new antennas, no new radios, no new site visits.
Privacy-Preserving by Design
All sensing inference runs at the edge. No raw RF data or personally identifiable information leaves the cell site. Federated learning keeps model improvements local — meeting GDPR, CCPA, and sovereign data requirements out of the box.
NVIDIA AERIAL PLATFORM
Built on NVIDIA Aerial
OranSense vRAN Next is validated on the NVIDIA Aerial SDK — the same GPU-accelerated software stack powering AI-RAN deployments with T-Mobile, Nokia, and leading MNOs worldwide as they build toward 6G.
GPU-Accelerated RAN
NVIDIA Aerial SDK runs L1/L2 baseband on CUDA GPUs — the same infrastructure OranSense vRAN Next is built on.
AI-Native 6G Foundation
NVIDIA and global telco leaders committed to build 6G on open, AI-native platforms. OranSense is aligned with this roadmap.
Omniverse Digital Twin
NVIDIA AODT (Aerial Omniverse Digital Twin) enables simulation-first network design — integrated with OranSense orchestration layer.
Edge AI Inference
CUDA-X AI microservices run on the same GPU fabric as the RAN, enabling sub-millisecond inference loops at the cell site.
RESOURCES
Platform Documentation
Ready to Transform Your RAN?
Talk to our team about deploying the OranSense Sensing Platform in your network. We work with MNOs, neutral hosts, and enterprise private network operators.