Reinforcement Learning-Based SCMA Codebook Design for Uplink Rayleigh Fading Channels

Sparse Code Multiple Access (SCMA) is a promising technique for next generation mobile communication systems. In this letter, the problems surrounding the design of an SCMA codebook problem are confronted through the use of Artificial Intelligence (AI) techniques. The suggested algorithm is based on Reinforcement Learning (RL). The design parameters include a set of actions, a set of states, and a reward function. It is shown that the proposed algorithm is capable of generating codebooks that include superior metric values and optimized signal constellations based on low levels of searching complexity. The low-complexity feature in the RL-based construction algorithm ensures that it is more suitable to applications that rely on large-scale SCMA schemes.