Massive Random Access With Trellis-Based Codes and Random Signatures

Abstract

We investigate unsourced and grant-free massive random access in which all the users employ the same codebook, and the basestation is only interested in decoding the distinct messages sent. To resolve the colliding user packets, a novel approach relying on user transmissions with random-like amplitudes selected from a large number of possible signatures is proposed. The scheme is combined with convolutional coding for error correction. The receiver operates by first identifying the signatures used by the transmitting nodes employing a sparsity-based detection algorithm, and then utilizing a trellis-based decoding algorithm. Despite its simplicity, the proposed solution offers excellent performance.