Dynamic Traffic Steering for Networked Robotics Using 3GPP-Compliant Application Functions

Low-latency communication in 5G networks enables advanced applications in networked robotics, such as collaborative visual Simultaneous Localization And Mapping (vSLAM), where robots share real-time sensor data to build accurate maps and localize themselves. To reduce onboard energy consumption, robots can offload compute-intensive tasks—such as feature extraction and map fusion—to nearby Multi-access Edge Computing (MEC) servers. However, existing 5G traffic steering mechanisms often ignore the dynamic compute load at MEC sites, leading to service congestion and degraded Quality of Service (QoS). This paper proposes a novel traffic steering scheme that jointly considers real-time network conditions and MEC compute availability to optimize service instance selection. Leveraging the 3GPP-defined Application Function (AF), our system dynamically associates mobile clients with the most suitable service endpoint, minimizing end-to-end latency and improving load distribution. We outline an experimental evaluation plan including key performance metrics, a 5G-enabled testbed, and representative robotic workloads. The proposed scheme is based on Release 18 and adaptable to future beyond-5G architectures.