Design of a Probabilistic Link Adaptation Model for Fifth Generation Millimeter-Wave Networks Using Adaptive Modulation and Coding and Channel Prediction to Minimize HARQ Retransmissions

The advancement of Fifth Generation millimeter-wave networks necessitates a reliable link adaptation model that is suitable to maintain optimal data rates despite fluctuating channel conditions. The conventional link adaptation model relies on average Signal-to-Interference-plus-Noise Ratio measurements that suffer from delayed and inaccurate modulation as well as coding selection of unpredictable channel variations, user mobility, and feedback latency. These models reject the probabilistic nature of fading and interference, resulting in suboptimal throughput, high packet error rates, and unstable performance in dynamic propagation environments. These limitations are addressed by the proposed Probabilistic Link Adaptation with Channel Prediction model by merging channel forecasting, reliability-aware throughput estimation, and adaptive recursive learning. It employs a probabilistic autoregressive channel model that predicts future channel coefficients, captures the mean and uncertainty of time-varying links. Further, the Gaussian-based Signal-to-Interference-plus-Noise Ratio distribution estimates packet error probability for each Modulation and Coding Scheme. The reliability-constrained optimization rule applies to attain an optimal Modulation and Coding Scheme that improves throughput. The experimental validation is undertaken by various measures and attains a higher prediction accuracy of 97.4 % and throughput of 99.8 Mbps, respectively.