Deep Learning Detectors with Pilot Interval Reduction in Uplink Non Orthogonal Multiple Access

DeepMuD: Multi-user Detection for Uplink Grant-Free NOMA IoT Networks via Deep Learning

“…3 There is no theoretical way to find/select the optimum number of layers or cells in each layer [18]. Thus, as being in all DL-aided communications applications [6]- [17], these parameters are empirically determined, such that increasing the sizes do not provide a noteworthy gain in learning performance and the network performance converges.…”

“…Therefore, the machine learning algorithms, particularly deep learning (DL), have been proposed in NOMA-involved systems for modulation, constellation design and resource allocation [8]- [10]. Besides, the detector designs 1 via DL instead of conventional detectors have also been investigated for basic downlink [12]- [15] and uplink [16], [17] NOMA schemes. However, none of the previous uplink designs [16], [17] considers a grant-free access where the number of devices is fixed to only two, which is very low for the IoT networks.…”

“…Besides, the detector designs 1 via DL instead of conventional detectors have also been investigated for basic downlink [12]- [15] and uplink [16], [17] NOMA schemes. However, none of the previous uplink designs [16], [17] considers a grant-free access where the number of devices is fixed to only two, which is very low for the IoT networks. Furthermore, in [16], three frames are required to detect one frame data to acquire the channel information of the devices.…”