COHERE -- Congestion-aware Offloading and Handover via Empirical RAT Evaluation for Multi-RAT Networks

Abstract

The evolution of wireless networks and radio access technologies (RATs) has transformed communication from user-driven traffic into a dynamic ecosystem of autonomous systems, including IoT devices, edge nodes, autonomous vehicles, AR/XR clients, and AI-powered agents. These systems exhibit diverse traffic patterns, latency requirements, and mobility behaviors, increasingly operating across overlapping heterogeneous RATs such as 5G, WiFi, satellite, NB-IoT, LoRaWAN, Zigbee, etc. This multi-RAT coexistence creates opportunities for intelligent access, mobility, and routing strategies. However, most mobility decisions still rely heavily on RSSI, which neglects RAT-specific features, congestion, queuing delays, and application needs, favoring high-power links over optimal ones. To address this gap, we propose chrome (Congestion-aware Offloading and Handover via Empirical RAT Evaluation), a multi criteria framework for dense multi-RAT networks. chrome enhances RSSI with multiple criteria and applies the Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS) to rank available RATs. Criteria weights are determined using both subjective (operator-driven) and objective (measurement-based) approaches. Based on this ranking, chrome performs intelligent cross-RAT offloading to reduce congestion on over-utilized links. We evaluate chrome in a dense SDN-controlled 5G/WiFi Multi-RAT environment using Mininet WiFi. Compared to RSSI-only handover, COHERE reduces the load on the congested RAT by up to 32%, reduces total handovers by 25%, lowers handovers to the congested RAT by 55%, and improves link delay by up to 166%, while maintaining comparable or up to 11% higher throughput. These results demonstrate that guarded, multi-criteria decision-making can exploit RAT coexistence to deliver robust, congestion-aware performance across heterogeneous deployments.