A New SS-VDS-SLM Technique for PAPR Reduction in High-Mobility Real-Time OTSM

This paper presents a novel semi-supervised vectored delay-sequency selective mapping (SS-VDS-SLM) technique designed to address the peak-to-average power ratio (PAPR) reduction challenge in orthogonal time sequency multiplexing (OTSM) systems. The proposed method leverages in-phase-quadrature delay-sequency symbol vectors and integrates a hybrid agglomerative hierarchical clustering with a weighted K-nearest neighbour algorithm to effectively reduce transmission computational complexity (CC) and transmission latency (TL) and enhance spectral efficiency (SE), while simultaneously mitigating out-of-band emissions (OOB). The approach ensures that the bit error rate (BER) and error vector magnitude (EVM) performance are preserved within the 3GPP doubly dispersive high-mobility delay Doppler multipath channels. Benchmark comparisons against ten contemporary PAPR reduction schemes in both OTSM, orthogonal frequency division multiplexing (OFDM) and both of their variants including the base orthogonal time frequency space (OTFS), demonstrate the superior performance of the SS-VDS-SLM technique in OTSM. Further, the proposed method offers a more effective trade-off among the PAPR, OOB, CC, TL, SE, BER, and EVM performance metrics compared to the ten considered contemporary schemes. Thus, the proposed SS-VDS-SLM technique provides a more efficient and scalable solution for reducing PAPR in high-mobility speeds of 500 km/h and 1000 km/h in real-time OTSM system applications.