Novel OTFS system designs for 6G communication networks

Conventional modulation methods, though effective in many scenarios, encounter difficulties when dealing with high-speed communication in dynamic environments. The rapid movement of devices, along with the impacts of multipath propagation and Doppler spread, can introduce complexities that degrade communication quality. To address these challenges, the orthogonal time frequency space (OTFS) waveform emerges as an innovative solution, bridging the gap between the demands of next-generation communication technologies and the intricacies of high-speed data exchange. OTFS represents a groundbreaking modulation technique that reimagines the approach to communication within the time-frequency domain. Unlike traditional methods, OTFS adopts a more comprehensive approach by simultaneously considering both time and frequency domains. However, there is still a need for the design of more sophisticated OTFS-based modulation schemes to meet the rigid requirements of the 6G standard. In this thesis, three performance improvement methods for OTFS are presented. In Chapter 3, a deep learning (DL)-based technique named autoencoder (AE)-based enhanced OTFS (AEE-OTFS) to improve the error performance of OTFS is proposed. An AE structure is exploited to learn a set of high-dimensional symbols and maximize the squared minimum Euclidean distance (SMED) between them. Unlike complex doubly-dispersive channels, this AE is trained under the simpler conditions of an additive white Gaussian noise (AWGN) channel. The encoder and decoder of AE are then repurposed as the mapper and demapper blocks in a real-time OTFS system. The approach also establishes a theoretical frame error rate (FER) upper bound. Simulation results demonstrate that AEE-OTFS outperforms conventional OTFS in terms of FER performance. The addition of a learned detector (LD) significantly reduces decoding complexity. These findings suggest that AEE-OTFS holds promise for certain 6G applications, particularly in high-mobility and flexible scenarios. In Chapter 4, a block-wise index modulation (IM) technique named joint delay-Doppler IM-based OTFS (JDDIM-OTFS) is designed to enhance the error performance of OTFS. In this chapter, a theoretical upper bound on the bit error rate (BER) is also established. Computer simulations reveal that JDDIM-OTFS surpasses conventional OTFS, OTFS-IM, and delay-IM OTFS (DeIM-OTFS) systems in terms of BER when combined with matched-filtered Gauss-Seidel (MFGS) detector with a maximum likelihood detector (MFGS-ML) and MFGS with a greedy detector (MFGS-GRD). The adoption of a Greedy detector within MFGS substantially reduces decoding complexity. JDDIM-OTFS exhibits remarkable error performance in scenarios characterized by high mobility and flexibility, making it an exciting choice for specific 6G use cases. Finally, in Chapter 5, for reducing the peak-to-average power ratio (PAPR) of OTFS, a method is suggested that utilizing unitary transformations such as Walsh-Hadamard transform (WHT), Zadoff-Chu transform (ZCT), and discrete cosine transform (DCT). The performance of this method is evaluated across various frame sizes and compared to clipping and discrete Fourier transform-spread (DFT-s OTFS) techniques. Additionally, the impact of this approach on BER performance is assessed against other methods. Computer simulations demonstrate that this proposed method significantly reduces PAPR while introducing a compromise of approximately 1.5 dB in error performance. The substantial PAPR reduction, coupled with the absence of the need for receiver-side information, highlights the significance of this approach.