I. Introduction
With the explosive growing demand on network capacity, throughput and connected wireless devices, the mobile broadband network is evolved into fifth generation, in which the enhanced mobile broadband (eMBB), ultra-reliable low latency communication (URLLC), and massive machine type of communication (mMTC) are three typical application scenarios. Current air interface technologies, such as orthogonal frequency-division multiple access (OFDMA) cannot fulfill the requirements in the above scenarios as orthogonal multiple access (OMA) assigns the time-frequency resources to each user exclusively. Therefore, OMA is spectrum inefficient and cannot support large throughput as well as the massive connected users in the network. In contrast, non-orthogonal multiple access (NOMA) where each resource unit shared by multiple users is supposed to be more spectrum efficient. Sparse code multiple access (SCMA) [1] is a code domain NOMA and is designed based on the multi-dimensional sparse signal constellation. Due to the shaping gain of multi-dimensional constellation, SCMA has a better bit-error-rate (BER) performance compared with other NOMA schemes, such as low density signature (LDS) [2].