Adaptive Traffic Steering for Cellular Networks: Leveraging Predictive QoS Metrics for Application-specific Optimization

<p>The rising demand of Cooperative, Connected, and Automated Mobility (CCAM) applications, which require strict network performance guarantees, has led to the integration of multiple access types in cellular network architectures. To this end, the Access Traffic Steering, Switching, and Splitting (ATSSS) framework incorporates non-3GPP access technologies into the 5G network. As these traffic steering rules often overlook application-specific needs, this paper introduces a 5G-compliant traffic steering architecture that integrates Predictive Qualityof-Service (PQoS) metrics to address both dynamic network conditions and strict, possibly changing application requirements. Through a programmable Smart-ATSSS module, the architecture formulates optimized steering rules, and its adaptability is demonstrated across different CCAM scenarios.</p>