Low-Complexity SCMA Detection for Unsupervised User Access

Non-orthogonal multiple access schemes (NOMA), such as sparse code (SCMA), are among the most promising technologies to support massive numbers of connected devices. Still, minimize transmission delay and maximize utilization channel, “grant-free” NOMA techniques required that eliminate any prior information exchange between users base-stations. However, if a large number transmit simultaneously in an “unsupervised” manner, (i.e., without signaling for controlling corresponding patterns), it is likely may share same frequency-resource element, rendering user detection impractical. In this context, we present new multi-user approach, which aims performance, with respect given processing latency limitations. We show our approach enables practical grant-free SCMA hundreds interfering users, complexity up two orders magnitude less than conventional approaches.