Affect-aware QoE-to-SLA Mapping for 5G-Advanced and 6G Network Slicing

Abstract

This paper proposes an Affect-to-Service Level Agreement (A2S) framework for affect-aware network slicing in 5G-Advanced and 6G networks. A2S estimates the user affective state in a normalised valence–arousal space and converts it into a Mean Opinion Score (MOS)-based Quality of Experience (QoE) objective. The resulting QoE requirement is mapped to Service Level Agreement (SLA) parameters: guaranteed bitrate (GBR), packet delay budget (PDB), packet error rate (PER), and scheduling weight. The framework combines affect-to-QoE translation and QoE-to-SLA mapping as a lightweight operator-domain extension around the Network Slice Selection Function (NSSF). A 5G-LENA/ns-3 urban macro evaluation with 150 user equipment units and mixed traffic is used to assess the framework. The results show that A2S improves average MOS from 3.92 to 4.11, increases 10th-percentile MOS from 3.18 to 3.57, and reduces 99th-percentile latency from 59.4 ms to 47.1 ms, while preserving Jain fairness above 0.85.