부산대학교 도서관

Orbital angular momentum with index modulation (OAM-IM) has a great potential of providing high spectral efficiency and energy efficiency by utilizing the indices of the orthogonal OAM modes. However, the harsh requirement of perfect alignment of the transceiver beams introduces great challenges to OAM-IM wireless communications. Therefore, we first propose an angle of arrival (AoA)-based robust detector for the misaligned OAM-IM system, which explicitly estimates the AoA of the OAM beam and then utilizes the estimate to detect the transmitted symbols. To further reduce the system overhead and complexity, we propose another deep learning (DL)-based robust detector, which implicitly estimates the AoA and directly recovers the transmitted information bits. By using the dataset collected through simulation, the first step is to train the DL-based robust detector offline to minimize the mean-squared error, and the second step is to use the trained model for real-time OAM-IM signal detection online. Numerical simulations validate that the both proposed robust detectors can address the channel distortion in OAM channels with beam misalignment and achieve superior bit error rate (BER) performance at high spectral efficiency. Moreover, the proposed DL-based robust detector is less complicated on runtime than the traditional OAM-IM detector.