PaperlayerView exact revision on arXiv

Source-linked AI summary

OTFS: A New Generation of Modulation Addressing the Challenges of 5G

Ronny Hadani, Anton Monk

arXiv:1802.02623v1cs.IT

TL;DR

The paper addresses 5G scenarios where OFDM faces complexity, channel selectivity, mobility, interference, and adaptation challenges. It introduces OTFS, which maps QAM symbols onto delay-Doppler waveforms with invariant, separable, and orthogonal channel coupling. The reported results include MIMO-order capacity scaling, mobility and interference robustness, and improved small-packet operation, with stated limitations involving receiver complexity and cyclic-prefix overhead.

  • Problem

    5G use cases require efficient multi-user MIMO and reliable operation under dynamic channels, interference, power constraints, and high-frequency phase noise beyond practical OFDM adaptation.

  • Method

    OTFS multiplexes QAM symbols on time-frequency-spread waveforms whose delay-Doppler interaction provides invariant, separable, and roughly orthogonal coupling, implementable with multicarrier preprocessing.

  • Results

    OTFS provides spectral-efficiency advantages in high-order MIMO, robustness to Doppler and interference, and a small-packet allocation that maximizes link budget while minimizing retransmissions.

  • Takeaways & Limitations

    OTFS supports coherent diversity use across mobility, interference-limited, MIMO, and power-constrained 5G communication scenarios.

  • Takeaways & Limitations

    Receiver performance-complexity tradeoffs remain constrained by detector iteration limits or lattice-basis recalculation, while increased subcarrier spacing can increase cyclic-prefix overhead.

Abstract

from arXiv · show

In this paper, we introduce a new 2D modulation scheme referred to as OTFS (Orthogonal Time Frequency & Space) that multiplexes information QAM symbols over new class of carrier waveforms that correspond to localized pulses in a signal representation called the delay-Doppler representation. OTFS constitutes a far reaching generalization of conventional time and frequency modulations such as TDM and FDM and, from a broader perspective, it establishes a conceptual link between Radar and communication. The OTFS waveforms couple with the wireless channel in a way that directly captures the underlying physics, yielding a high-resolution delay-Doppler Radar image of the constituent reflectors. As a result, the time-frequency selective channel is converted into an invariant, separable and orthogonal interaction, where all received QAM symbols experience the same localized impairment and all the delay-Doppler diversity branches are coherently combined. The high resolution delay-Doppler separation of the reflectors enables OTFS to approach channel capacity with optimal performance-complexity tradeoff through linear scaling of spectral efficiency with the MIMO order and robustness to Doppler and multipath channel conditions. OTFS is an enabler for realizing the full promise of MUMIMO gains even in challenging 5G deployment settings where adaptation is unrealistic.

1. OTFS – A NEXT GENERATION MODULATION

OTFS is proposed as a next-generation 2D modulation scheme for diverse 5G scenarios where legacy OFDM faces complexity, selectivity, and adaptation challenges. It combines spreading and orthogonality through delay-Doppler waveforms to improve robustness, efficiency, and performance-complexity tradeoffs across key use cases.

  • Motivation: 5G multi-user MIMO and dynamic channels challenge OFDM because practical architectures cannot achieve capacity efficiently and adaptation can become unrealistic.The paper motivates eliminating adaptation while preserving performance under diverse deployment conditions.
  • OTFS approach: OTFS spreads QAM-carrying waveforms across time and frequency while remaining roughly orthogonal under general delay-Doppler channel impairments.Its reflector interaction induces a simple, symmetric coupling between channel response and information symbols.
  • OTFS approach: OTFS combines spread-spectrum reliability and robustness with narrowband transmission’s high spectral efficiency and low complexity.This design fuses spreading for interference resilience and diversity gain with orthogonality for simpler channel coupling.
  • Enhanced Mobile Broadband (eMBB): For eMBB, OTFS enables spectral-efficiency scaling with increased MIMO order under any channel condition while supporting delay-Doppler equalization and MU-MIMO precoding.The paper contrasts this intrinsic delay-Doppler advantage with conventional time-frequency processing.
  • Internet of Things (IoT): For IoT, OTFS’s small-packet mode maximizes link budget and minimizes retransmissions under power and latency constraints, extending battery life and coverage.The transmit signal uses low PAPR, maximum available duration, and full time-frequency diversity.
  • High mobility: Under mobility, OTFS maximizes throughput, reliability, and performance consistency by treating Doppler as diversity while avoiding severe intercarrier interference.The paper identifies vehicle-to-vehicle and high-speed-train communication as representative scenarios.
  • URLLC and mm-Wave: For URLLC, OTFS is resilient to narrowband interference, enabling coexistence with URLLC packets and other narrowband interference.The paper also discusses phase-noise mitigation for potential mm-Wave communication without sacrificing capacity.

2. OTFS PRINCIPLES

OTFS generalizes time- and frequency-based modulation by placing QAM symbols on localized delay-Doppler waveforms whose channel interaction reflects reflector physics. This produces invariant, separable, and approximately orthogonal symbol coupling while linking communication processing with Radar-style delay-Doppler representation.

  • Signal representation: OTFS multiplexes QAM symbols over localized pulses in the delay-Doppler representation, generalizing conventional TDM and FDM schemes.The delay-Doppler representation uses delay and Doppler as its two coordinates and extends time and frequency representations.
  • Channel coupling: The OTFS channel interaction yields a high-resolution delay-Doppler image of wireless reflectors and a simple, symmetric coupling with information symbols.The resulting symmetry is characterized by invariance, separability, and orthogonality.
  • Channel coupling: Invariance gives every QAM symbol the same channel coupling, while separability exposes the channel’s distinct delay-Doppler diversity paths to each symbol.These properties prevent location-dependent coupling and destructive addition of separated reflections at the QAM-symbol level.
  • Channel coupling: Orthogonality confines received echoes to a small delay-Doppler region, allowing geometrically separated transmit pulses to remain roughly orthogonal at the receiver.The echo region is determined by the channel’s delay and Doppler spread and is smaller than the outer periods.
  • Modulation framework: The OTFS family smoothly interpolates between TDM and FDM, which arise as limiting cases when the delay or Doppler period approaches infinity.Different OTFS schemes correspond to different delay-Doppler representations parameterized by periods satisfying τ$ ∙ υ$ = 1.
  • Multicarrier interpretation: OTFS can overlay a conventional multicarrier transceiver by transforming delay-Doppler symbols to a reciprocal time-frequency grid before multicarrier transmission.The implementation uses an SFFT followed by conventional IFFT-based multicarrier transmission, with corresponding post-processing at reception.

EQUALIZATION AND PRECODING

The section compares time-frequency and delay-Doppler equalization and precoding for MU-MIMO. OTFS uses invariant, separable, orthogonal coupling to reduce condition numbers and improve receiver complexity and downlink SNR.

  • Equalization: In multicarrier MU-MIMO, QAM symbols are multiplexed over time-frequency grid points, creating parallel local MIMO channels whose condition numbers vary with channel selectivity.Sphere-detector convergence depends critically on each local channel's condition number.
  • Equalization: High condition numbers slow sphere-detector convergence, causing complexity that can grow exponentially with the MIMO order.Time-frequency selectivity can make many local autocorrelation matrices ill-conditioned.
  • Equalization: Limiting detector iterations sacrifices performance, while lattice reduction preserves performance but requires recomputing reduced bases across coherence intervals.Commercial MIMO implementations commonly use reduced-complexity detectors, and full sphere detection is generally limited to four spatial streams.
  • Equalization: OTFS multiplexes QAM symbols on the delay-Doppler grid, where invariant, separable, orthogonal coupling produces a global channel matrix with a lower condition number through averaging.The delay-Doppler formulation is analyzed under simplified assumptions relating its autocorrelation matrix to the average of local time-frequency matrices.
  • Equalization: A simulation over Gaussian and realistic 4x4 MIMO channels found significantly lower average delay-Doppler than time-frequency condition numbers, implying superior OTFS spatial multiplexing.The realistic channel used 30 km/h maximum Doppler spread and 3 microseconds maximum delay spread over a 20 MHz, 10-second region.
  • Precoding: In downlink precoding, time-frequency selectivity can create fades that increase normalization terms and degrade SNR, making non-regularized zero-forcing strictly sub-optimal.Regularized-inverse variants can provide small SNR improvements, but the discussion restricts attention to non-regularized zero-forcing.
  • Precoding: For OTFS downlink precoding, averaging roughly independent local channel matrices lowers the global condition number, enabling higher-SNR ZF THP close to downlink capacity.Lattice reduction can improve SNR further without compromising complexity because the reduced basis need only be computed once per frame.

4. OTFS PERFORMANCE ADVANTAGES OVER OFDM

This section compares OTFS with OFDM across five 5G use cases, emphasizing performance-complexity tradeoffs under mobility, interference, and challenging propagation conditions. Simulations show pronounced OTFS gains for higher-order MIMO and small packets, including a 36%–53% gap for 4x4 MIMO at about 19 dB SNR.

  • 5G use cases: OTFS is evaluated against OFDM for eMBB, high-mobility communication, IoT, URLLC coexistence, and mm-Wave communication.The study characterizes each use case's objective and technical challenge, then supports the theoretical comparison with simulations.
  • Enhanced Mobile Broadband: The eMBB evaluation targets spectral-efficiency scaling with MIMO order while maintaining an optimal performance-complexity tradeoff.The comparison uses 3GPP LTE numerology and assumptions, with OTFS Turbo and OFDM maximum-likelihood receivers.
  • Evaluation setup: The evaluation covers a 10 MHz system with a 3GPP TDL-C channel, 1 msec TTI, ideal channel estimation, LTE Turbo coding, and OTFS Turbo versus OFDM-ML reception.These settings define the principal simulation assumptions for the large-packet comparison.
  • Equalization results: 36%–53%: at about 19 dB SNR, OTFS exceeds OFDM for 4x4 MIMO, depending on the OFDM receiver.For large packets at 30 km/h, the reported gain is attributed to the channel condition-number effect rather than additional spreading diversity.
  • Packet-size robustness: OTFS maintains performance across packet sizes because every QAM symbol experiences the channel's full diversity, unlike small OFDM packets exposed to selective fades.This produces more consistent throughput, while small OFDM packets may become trapped in deep time- or frequency-selective fading and rely on FEC recovery.

4 PRB

OTFS addresses packet-size, mobility, interference, and power-constraint challenges by spreading symbols across delay-Doppler diversity modes. The reported results show gains in consistency, SNR, BLER, link budget, and retransmission requirements over conventional multicarrier approaches.

  • 4 PRB: 99% of OTFS users achieve more than 12dB SNR, compared with around 50% of OFDM users.The authors also report that 90% of OTFS users have more than 10dB SNR gain compared with 90% of OFDM users.
  • 4 PRB: At 120 km/h, OTFS outperforms OFDM at 10% coded BLER by 2.4 dB to 4 dB for large packets.The comparison uses 300 ns RMS delay spread, 2x2 MIMO, and 16 QAM and 64 QAM.
  • 4 PRB: At 500 km/h, OTFS with 15 kHz SCS outperforms OFDM with 60 kHz SCS by about 2.6 dB.Using smaller SCS also avoids the proportional cyclic-prefix overhead increase caused by shortening OFDM symbols.
  • 4 PRB: OTFS spreads each QAM symbol over the full time-frequency grid, making diversity gain independent of packet size and improving throughput consistency.This contrasts with time-frequency allocation, where small packets can fall into deep fades when channel adaptation is unavailable.
  • 4 PRB: In OTFS, URLLC interference is spread over the delay-Doppler grid before FEC decoding, producing only a small SNR degradation rather than total decoding confusion.The cited comparison describes narrowband interference as directly disrupting multicarrier data bits.
  • 4 PRB: Doppler transversal allocation combines low PAPR and maximum transmit duration with full time-frequency diversity, overcoming multicarrier capacity saturation.The cited limitation is that fixed QAM order prevents multicarrier modulation from simultaneously maintaining low PAPR and extracting diversity gain.
  • 4 PRB: At a 1% packet error rate, OTFS requires 7dB less transmission power than hopped SC-FDMA for a 168-symbol QPSK packet.The simulation uses a single-antenna, single-PRB packet through a frequency-selective channel at code rate 0.9.

5. SUMMARY

OTFS multiplexes QAM symbols over localized delay-Doppler pulses and links wireless communication with radar-like channel representation. The paper reports invariant channel interaction, MIMO-scaled capacity, interference resilience, mobility robustness, and power-efficient small-packet transmission.

  • 5. SUMMARY: OTFS multiplexes QAM information symbols over localized pulses in the delay-Doppler representation.The paper presents OTFS as a generalization of TDMA and OFDM.
  • 5. SUMMARY: OTFS waveforms capture wireless reflector physics, yielding a high-resolution delay-Doppler radar image and invariant, separable, orthogonal channel interaction.The resulting interaction gives received QAM symbols a common localized impairment and coherently combines delay-Doppler diversity branches.
  • 5. SUMMARY: OTFS channel-symbol coupling enables capacity to scale linearly with MIMO order while maintaining an optimal performance-complexity tradeoff.The paper reports spectral-efficiency advantages over traditional modulation schemes including OFDM under general channel conditions.
  • 5. SUMMARY: OTFS supports URLLC overlays through interference resilience and supports mobility through delay-Doppler diversity gain.These properties are identified as suitable for ultra-reliable low-latency packets and communication under mobility conditions.
  • 5. SUMMARY: Doppler transversal allocation targets low-PAPR, power-constrained small-packet transmission while extracting full time-frequency diversity and maximizing restricted capacity.The paper reports superiority over SC-FDMA and hopped SC-FDMA for IoT applications.
  • 5. SUMMARY: OTFS is presented as suited to eMBB deployments requiring MIMO-scaled capacity, high spectral efficiency, and reliability under diverse channel conditions.The stated application context includes 3GPP deployment scenarios focused on MU-MIMO and high Doppler environments.
Loading 1802.02623v1…