Risk Sensitive Semantic Flow Management in Capacity Constrained Wireless Systems Kapasite Kisitli Kablosuz Haberleşme Sistemlerinde Risk Duyarli Anlamsal Veri Akişi Yönetimi
34th Signal Processing and Communications Applications Conference, SIU 2026, İstanbul, Türkiye, 7 - 10 Temmuz 2026, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/siu71813.2026.11636807
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Anahtar Kelimeler: 6G, artificial intelligence, Reinforcement learning, semantic communication, wireless communication
- Ankara Yıldırım Beyazıt Üniversitesi Adresli: Evet
Özet
This study addresses the capacity constraints created by increasing device and data density using an adaptive data management approach based on semantic communication. In a fixed-capacity system with multiple data streams, the primary objective is to protect critical data and reduce unnecessary data load. To this end, a risk-sensitive reinforcement learning-based mode adaptation method is proposed. The system consists of three data types: general user traffic, sensor data, and camera streams. While critical sensor data is prioritized for protection, non-critical sensor and camera streams are dynamically scaled through a total of nine modes with three levels of adaptation. The decision-making mechanism selects the most suitable mode based on traffic density and risk level. Simulation results show that, especially under high traffic loads, the proposed method significantly reduces average data loss and increases system efficiency compared to the fixed-mode approach.