Code-Domain Collision Resolution Grant-Free Random Access for Massive Access in IoT

Code domain grant-free random access (GFRA) is regarded as a potential framework to serve massive access in Internet of Things (IoT). In this paper, we propose an LT -collision resolution GFRA ( LT -GFRA) scheme by combining a pilot set containing L orthogonal pilots and a T -order codebook, where each activated user equipment (UE) randomly selects one of L pilots, and directly sends to the base station (BS) followed with data encoded by the T -order codebook together. Thus, the BS can receive L different frames and a conventional collision occurs when more than one UE select the same pilot. Moreover, we design a successive cancellation then joint decoding (SCJD) decoder, and prove the BS can decode at most T UEs from the frame on the same pilot and can recover up to LT UEs. Then, we derive the decoding error probability of our LT -GFRA scheme in Rayleigh fading channel, and further derive the access failure probability (AFP) and the system throughput in the single- and multiple-antenna systems. We also extend and evaluate our LT -GFRA scheme in the shadowed-Rician fading channel. Finally, simulations validate our analytical results, and indicate that our LT -GFRA scheme can greatly outperform the state-of-art schemes for massive access in IoT.