Rate Grouping Based Non-Orthogonal Random Access Aided by Online Traffic Load Prediction

With user collision being the main concern for massive random access, non-orthogonal multiple access (NOMA) is an important tool to address the issue because it allows users to share the same channel resources. However, more refined access control schemes are expected in consideration of the explosively increasing traffic in modern communication systems. In this paper, we aim to maximize the access ratio with the aid of traffic load prediction and NOMA technique. The user equipments (UEs) are divided into groups according to their rate requirements. Given the grouping policy, the overall access success probability is analyzed based on Shannon capacity limits. A dynamic programming-based resource grouping algorithm is proposed to optimize the grouping policy of resources. The grouping optimization procedure depends on the output of a knowledge-driven online traffic load predictor, which gives the specific prediction of the traffic in each rate group. The proposed scheme is user-friendly because the prediction and the grouping optimization are both carried out at the base station (BS). Numerical results proved the effectiveness and superiority of the proposed scheme over non-grouping schemes.