Agentic data science with the Data Agent Kit
cloud.google.com·6h ago
AI Research
arXiv CS.AI·Aug 26·5 min read
Place, Slice and Schedule: Hierarchical O-RAN Control of a Tethered mmWave UAV-gNB
TL;DR
A hierarchical control system for tethered UAV-mounted 5G base stations was developed, enhancing resource management and improving service delivery metrics significantly. The system achieved up to 42% better on-time delivery for ultra-reliable low-latency communication (URLLC) compared to traditional methods.
✦ Why It Matters
Consider implementing hierarchical control systems for UAV-based networks to optimize resource allocation and improve service delivery.
Key Takeaways
- Explore hierarchical control frameworks for UAV networks.
- Utilize O-RAN for better resource management.
- Implement DeepSets Soft Actor-Critic for user scheduling.
- Aim for improved SLA satisfaction and delivery metrics.
How It Works
The system employs a two-tier control mechanism where a Non-Real-Time RIC rApp manages the overall placement of the UAV and allocates resources based on aggregated performance indicators. In parallel, a Near-Real-Time RIC xApp dynamically schedules resources for individual users using a DeepSets Soft Actor-Critic model, which treats users as an unordered set, allowing for efficient resource allocation that adapts to real-time conditions.