Enriched HARQ Feedback for Link Adaptation in 6G: Optimizing Uplink Overhead for Enhanced Downlink Spectral Efficiency

Conventional hybrid automatic repeat request (HARQ) feedback uses a single-bit acknowledgment, ensuring reliability but lacking the detail required for effective link adaptation (LA) during retransmissions. To address this, we introduce an enriched HARQ feedback (EHF) framework that goes beyond reliability gains to enhance spectral efficiency by reporting delta adjustments in modulation and coding scheme (MCS) levels from user equipment (UE) to the gNodeB (gNB). This framework features two innovations. First, our learning-driven adaptive quantization (LAQ) employs a-priori statistics to refine delta MCS quantization within fixed-size UE feedback. By tailoring quantization levels to channel-condition–based clusters, LAQ enhances downlink (DL) performance over existing techniques and provides strong out-of-distribution generalization. Second, we extend the EHF framework with a flexible variant where feedback bit size is optimized using a novel LA metric to boost DL performance and minimize uplink (UL) overhead, overcoming the rigidity of fixed-size feedback. Simulation results in a dense urban-macro deployment with file transfer protocol model 3 traffic show gains in DL indicators, including average physical resource block utilization (APU). EHF with LAQ improves DL APU by about 22% over conventional HARQ with Chase combining (CC) at 1-bit resolution. The flexible variant reduces average UL overhead by 1 bit in clusterless feedback for comparable performance and favors LAQ by 23% in end-user experienced DL throughput over CC. These findings underscore the effectiveness of LAQ and our optimized feedback metric in improving retransmission efficiency and feedback in HARQ systems. Complexity analysis supports the feasibility of both gNB and UE-side operations for deployment.