R1-2508471
discussion
Discussion on waveform for 6GR air interface
From Thales
Summary
Thales presents a comparative waveform analysis for 6G NTN, evaluating CP-OFDM, DFT-s-OFDM, WOLA, UFMC, BF-OFDM, F-OFDM, and OTFS against satellite link impairments including HPA non-linearities, Doppler, and delay. The document contains 11 observations and 3 proposals focusing on PAPR, ACLR, BER, and multi-user multiplexing performance, ultimately advocating for DFT-s-OFDM downlink study and PAPR reduction techniques.
Position
Thales presents technical case for DFT-s-OFDM as the superior waveform for NTN downlink based on PAPR reduction and ACLR performance under HPA non-linearities, showing 2.5–3 dB PAPR reduction over CP-OFDM for single-user single-beam. They demonstrate that DFT-s-OFDM's PAPR benefit diminishes to less than 1 dB when more than five beams are active in SDMA, and that PTS-enhanced DFT-s-OFDM fails to sustain PAPR improvement over multiple beams due to inability to operate PTS across beams. They propose studying DFT-s-OFDM in downlink for NTN-based 6G radio access and investigating candidate waveform performance under carrier frequency and time offsets for GNSS-free physical layer operation. They highlight that filtered waveforms (UFMC, BF-OFDM, WOLA, F-OFDM) lose out-of-band emission advantages after HPA amplification, with OBO improvements over CP-OFDM rarely exceeding 0.5 dB.
Key proposals
- Observation 1 (Sec 3.3.1): DFT-s-OFDM achieves a reduction of approximately 2.5 to 3 dB across the waveform distribution compared to baseline CP-OFDM in terms of PAPR for downlink transmission, while other evaluated waveforms demonstrate comparable PAPR values with variations of less than 1 dB.
- Observation 2 (Sec 3.3.1): Before amplification, UFMC and BF-OFDM achieve out-of-band emissions about 40 dB below CP-OFDM, while WOLA and F-OFDM perform even better; after amplification, HPA-induced distortion limits out-of-band emission improvements across all waveforms.
- Observation 3 (Sec 3.3.1): DFT-s-OFDM consistently delivers superior ACLR performance due to lower PAPR and reduced susceptibility to PA-induced non-linear distortion; the out-of-band emission advantages of filtered waveforms are mostly lost after amplification, with OBO improvements over CP-OFDM rarely exceeding 0.5 dB.
- Observation 4 (Sec 3.3.1): Phase noise negatively impacts all waveforms, but OTFS demonstrates superior robustness due to its time-domain symbol mixing which disrupts phase noise correlation, while UFMC experiences a slight additional degradation of less than 1 dB compared to others.
- Observation 5 (Sec 3.3.1): Under TDL channel with HPA non-linearities, OTFS outperforms all other waveforms in BLER due to sparse channel and effective MPA receiver operation, while UFMC is more vulnerable to ISI and ICI due to lack of a cyclic prefix.
- Observation 6 (Sec 3.3.1): UFMC demonstrates increased sensitivity to frequency synchronization errors with higher BER attributable to ISI from filter dispersion beyond its guard interval, compounded by inherent noise enhancement effects.
- Observation 7 (Sec 3.3.1): OTFS achieves stable BER across time offsets as its longer, slot-based CP ensures robust protection provided offsets remain within the CP duration, maintaining SNR without additional overhead compared to symbol-wise CP use in OFDM.
- Observation 8 (Sec 3.3.2.2): DFT-s-OFDM achieves a 3 dB PAPR reduction compared to CP-OFDM for a single beam, but this advantage diminishes to less than 1 dB when more than five beams are active.
- Observation 11 (Sec 3.3.2.5): PTS-enhanced DFT-s-OFDM scheme manages to maintain a roughly constant PAPR with the number of users in the frequency domain for a single beam but fails to sustain it over more than one beam.
- Proposal 1 (Sec 3.3.1): RAN1 to investigate the performance of candidate waveforms under varying carrier frequency and time offsets inherent to satellite links, specifically considering scenarios with significant uncertainty in the UE's position, focusing on conditions relevant to GNSS-free physical layer operation.
- Proposal 2 (Sec 3.3.2.5): RAN1 to study the performance of DFT-s-OFDM in downlink for non-terrestrial network (NTN)-based 6G radio access.
- Proposal 3 (Sec 3.3.2.5): As part of 6G study, for non-terrestrial network (NTN)-based 6G radio access, RAN1 to consider the evaluation and development of PAPR reduction techniques for both CP-OFDM and DFT-s-OFDM in the downlink.