Joint User and Data Detection in Grant-Free NOMA with Attention-based BiLSTM Network

We consider the multi-user detection (MUD) problem in uplink grant-free non-orthogonal multiple access (NOMA), where point has to identify total number and correct identity of active Internet Things (IoT) devices decode their transmitted data. assume that IoT use complex spreading sequences transmit information a random-access manner following burst-sparsity model, some data adjacent time slots with high probability, while others only once during frame. Exploiting temporal correlation, we propose an attention-based bidirectional long short-term memory (BiLSTM) network solve MUD problem. The BiLSTM creates pattern device activation history using forward reverse pass LSTMs, whereas attention mechanism provides essential context points. By doing so, hierarchical pathway is followed for detecting scenario. Then, by utilising sequences, blind estimated performed. proposed framework does not require prior knowledge sparsity levels channels performing MUD. results show achieves better performance compared existing benchmark schemes.