Three pillars. One governed inference layer.
InfraSense addresses two gaps the AI-RAN ecosystem currently has no standard answer for: guaranteed AI inference latency in the RAN, and operator-grade management for ISAC sensing.
Built on the NVIDIA AI Platform
OranSense solutions are purpose-engineered on NVIDIA's full-stack AI infrastructure — from NVIDIA Aerial for AI-RAN to NVIDIA Omniverse for digital twins and NVIDIA AI Factory for inference at scale.
Dedicated Inference Layer
Latency-Guaranteed AI Placement with Explicit Refusal
Every inference request declares a latency class (RT <10ms, Near-RT 10ms–1s, Non-RT >1s). The orchestrator places it only on infrastructure tiers capable of honoring that class. If no compliant placement exists, the request is explicitly refused rather than silently misrouted. Placement considers live infrastructure metrics, request priority, and an accuracy objective — selecting among model variants (fast / balanced / accurate) within the latency bound. Each run emits an auditable control action over E2, O1, or O2.
- Three enforced latency classes: RT (<10ms), Near-RT (10ms–1s), Non-RT (>1s)
- Hard tier rule: RT class admits only DU-adjacent / RT-RIC-adjacent nodes
- Explicit refusal when no compliant placement exists — no silent degradation
- Auditable control actions: beamforming, power, slice, handoff, sensing-schedule, O-Cloud resource recommendation
A1 SensingIntent Policies
Intent-Driven ISAC — Declare What the Network Should Sense
A new A1 policy type carries a sensing intent for a geographic/network scope: target metrics (e.g. localization accuracy, sensing coverage), resource constraints (ceiling on radio resources or transmit power spent on sensing), sensing mode, validity time, and priority. An rApp at the Non-RT RIC generates the policy; the Near-RT RIC derives concrete sensing configuration with constraint caps applied; a closed loop compares delivered performance against targets and retunes. The full chain runs live in the demonstrator.
- New A1 policy type: sensing intent with target metrics, resource constraints, sensing mode, validity time, priority
- rApp at Non-RT RIC generates policy; Near-RT RIC derives configuration with caps applied
- Live demonstrator: 0.1m critical tracking intent and 90% coverage intent running end-to-end
- Attaches to existing O-RAN A1 interface — adds a policy type, does not replace the stack
SMO Global Agent
Priority-Based Conflict Arbitration at the SMO Layer
A Global Agent at the SMO watches aggregate commitments across sensing intents and rApp/xApp resource consumers. When the sum over-commits the network, it resolves by priority: critical intents protected in full, lower-priority intents reduced to declared floors, non-sensing consumers trimmed proportionally — with every resolution reported for audit. Where no complete resolution exists, infeasibility is reported explicitly. The demonstrator reproduces a 170% over-commitment scenario and resolves it live.
- Watches aggregate commitments across all sensing intents and rApp/xApp consumers
- Priority resolution: critical intents protected; lower-priority reduced to declared floors; non-sensing trimmed proportionally
- Live demonstrator: 170% over-commitment scenario reproduced and resolved with auditable output
- Explicit infeasibility reporting — no papering over unresolvable conflicts
AODT — Digital Twin
A Living Model of Your Network
The Autonomous Operations Digital Twin (AODT) creates a continuously updated virtual replica of the physical RAN environment. Operators can simulate configuration changes, stress-test edge AI models, and validate new xApps before live deployment.
- Real-time synchronisation with live RAN telemetry
- What-if simulation for configuration and capacity planning
- xApp validation sandbox before production rollout
- NVIDIA Omniverse integration for 3D city-scale visualisation
Artifact Catalog
Governed Repository of Inference Entities
The Artifact Catalog is a governed repository of pre-trained sensing models, xApps, and inference pipelines — the source from which the inference layer selects placement candidates. Operators and enterprise customers can browse, license, and deploy on any O-RAN-compliant infrastructure.
- Inference entity catalog: the orchestrator selects model variants (fast / balanced / accurate) from this catalog
- Versioned xApp packages with automated compatibility checks
- Pre-trained models for crowd analytics, health monitoring, and security
- One-click deployment to OranSense edge nodes
Find the right solution for your network
Our solutions engineers will map InfraSense capabilities to your specific deployment scenario.
Talk to an Engineer