The Internet of Things (IoT) framework requires a massive number of connection thus demanding spectral efficient solutions such as Non-Orthogonal Multiple Access (NOMA). However, the main drawback of NOMA with successive interference canceler (SIC)-based detectors is the error floor in the uplink. In this letter, a reliable multi-user detection in uplink IoT NOMA is guaranteed by a Joint Maximum-Likelihood (JML) detector (i.e., optimum detection algorithm). We derive a closed-form upper bound of bit error rate (BER) of JML over Rayleigh fading channels for arbitrary number of IoT devices and an adaptive M-ary phase shift keying (M-PSK). Based on the extensive simulations, the derived expressions are validated and it is revealed that the JML improves the error performance in uplink NOMA and removes the error floor. Furthermore, regardless of the number of the IoT devices and modulation order, a full diversity order (i.e., number of receiving antennas) is guaranteed for each device.

中文翻译：

---

上行链路 NOMA-IoT 网络中的多用户联合最大似然检测：消除误差层

物联网 (IoT) 框架需要大量连接，因此需要频谱高效的解决方案，例如非正交多址 (NOMA)。然而，基于连续干扰消除器 (SIC) 的检测器的 NOMA 的主要缺点是上行链路中的错误本底。在这封信中，联合最大似然 (JML) 检测器（即最佳检测算法）保证了上行链路 IoT NOMA 中可靠的多用户检测。我们针对任意数量的物联网设备和自适应 M-ary 相移键控 (M-PSK)，推导出了 JML 在瑞利衰落信道上的比特误码率 (BER) 的闭合形式上限。基于大量的模拟，推导出的表达式得到验证，结果表明 JML 提高了上行 NOMA 中的错误性能并消除了错误底线。此外，