Skip to content

This work was supported by a grant of the Ministry of Research, Innovation and Digitization, CNCS – UEFISCDI, project number PN-IV-P2-2.1-TE-2023-1476, within PNCDI IV.


The Radio Spectrum Challenge

Exponential Growth of Wireles Comunication
This rapid expansion brings critical challenges that require attention:
  • Spectrum Congestion: Overlapping signals lead to reduced reliability and increased latency across dense IoT deployments
  • Persistent Interference: Uncontrolled interference degrades performance and can lead to service disruptions in mission-critical applications
  • Inefficient Resource Allocation: Current static allocation methods struggle to adapt to dynamic changes in network density and traffic patterns
  • Security Vulnerabilities:: A crowded spectrum increases opportunities for malicious attacks and unauthorized access
IoT Networks

IoT Networks

Massive Scale Connectivity – 41B devices by 2034

5G Networks

5G Networks

High-Speed Mobile Communications – mmWave & Network Slicing

6G Future Systems

6G Future Systems

AI-Native Networks for next-generation applications

Defence

Defence

Military & Critical Infrastructure – Radar & Tactical Communications

Security

Security

Threat Detection & Mitigation – Real-Time Defense

Use Case I:
LoRaWAN IoT Networks

Key Challenges

  • Cross-technology interference and signal degradation
  • Duty-cycle constraints within regulated frequency bands
  • Capture effect in high-density deployments


AI Solutions Delivered

  • LoRa spreading factor detection achieving ≈95% recall accuracy
  • Predictive traffic analytics enabling adaptive duty cycling


Impact: Enhanced Packet Delivery Ratio (PDR) and significant reduction of collision rates in dense LoRaWAN deployments — ensuring reliable IoT connectivity across complex urban environments.

Use Case II: 5G and B5G
Network Orchestration

  • 5G Requirements
5G networks demand high agility with mmWave frequencies, massive MIMO arrays, and advanced beamforming techniques requiring intelligent spectrum orchestration and adaptive resource management.
  • AI-Powered Classification
Our research implementations achieved 97% accuracy in 4G/5G waveform classification, establishing robust foundations for B5G network intelligence.
  • Anomaly Detection
Advanced Autoencoder (AAE) and PredNet algorithms for real-time threat identification and network anomaly detection
  • Jamming Mitigation
Intelligent detection with adaptive countermeasures and dynamic spectrum reallocation
  • Service Optimization
Predictive allocation for URLLC and eMBB service requirements

Use case III: Security & Resilience

Threat Landscape

Threat Landscape

• Coordinated jamming campaigns
• Network misconfigurations
• Protocol-level vulnerabilities

Real-Time Detection

Real-Time Detection

Advanced anomaly detection algorithms with millisecond response times

Dynamic Protection

Dynamic Protection

Adaptive frequency guard bands and intelligent beamforming countermeasures