Issa Chihaoui

Summary: In this paper, we propose two downlink multiple access architectures for networks where human‐type communication users (HTCUs) and machine‐type communication devices (MTCDs) coexist. The proposed schemes combine non‐orthogonal multiple access (NOMA), orthogonal frequency division multiplexing (OFDM), and index modulation (OFDM‐IM) concepts. In the first scheme, the base station (BS) transmits bits of HTCUs using modulated symbols and bits of MTCDs by the subcarrier activation pattern (SAP). This approach called IM‐NOMA with null subcarriers (IM‐NOMA‐NS) ensures that inactive subcarriers are always null, which improves the system bit error rate (BER) performance. To improve the spectral efficiency (SE), we propose a second approach, termed IM‐NOMA with dual‐mode modulation (IM‐NOMA‐DM), in which the HTCUs' bits are transmitted using two‐dimensional modulation and the MTCDs' bits are transmitted using one‐dimensional modulation and the SAP. For each proposed system, a near‐optimal low‐complexity detector, based on the energy‐detection (ED) and the log‐likelihood ratio (LLR) criterion, is provided to mitigate the detection burden of the optimal maximum‐likelihood (ML) detector. The BER performances and SE of the proposed schemes are investigated. The average BERs of IM‐NOMA‐NS and IM‐NOMA‐DM are derived in closed‐form expressions corroborated by the simulation results. We have proved numerically that the proposed schemes achieve a good trade‐off between BER performance, SE, and the number of supported users, making them more suitable for Internet of Things (IoT) applications.

In this paper we present a likelihood ascent search (LAS) detector for massive multiple-input multiple-output (MIMO) systems with unknown channel state information (CSI). In contrast to previous works on LAS detection that assumed perfect CSI available at the receiver, our architecture takes into account the channel uncertainty and is based on the knowledge of the channel distribution information (CDI) at the receiver. The proposed LAS detector does not explicitly use the estimated channel matrix: it jointly processes the received training sequence and data symbols to estimate the transmitted data vectors. First, the maximum-likelihood (ML) detection metric is derived for the proposed LAS detector. Then, a low complexity implementation approach, suitable for large MIMO systems, is proposed. The performances of the proposed architecture are assessed and compared to those of conventional LAS detectors.