Preamble Split Transmission and Joint Active User Detection for Massive Connectivity.

Low-power delay-tolerable services are very important applications of future large-scale IoT. Active user detection is its critical challenge. By exploiting the sparsity of the received signal, this challenge could be converted to a compressive sensing problem and hence solved by the approximated message passing (AMP) algorithm. In order to improve the active user detection performance, a preamble split transmission (PST) random access (RA) scheme is proposed, in which each partition of the preamble is sent in different coherent time duration to achieve the time diversity. Correspondingly, a joint active user detection (JAUD) algorithm is proposed to jointly detect the distributed split preambles at the base station. Simulation results show that the proposed access scheme achieves a higher detection accuracy at the cost of a longer access delay. According to the delay-tolerable characteristic of the target network, this cost could be acceptable.