OFDM vs OTFS Radar: A Quantitative Comparison

A Head-to-Head Comparison

Sections 1-3 established the ambiguity-function framework and derived the key resolution theorems. This section compares OTFS and OFDM as radar waveforms head-to-head across the relevant performance metrics: range resolution, velocity resolution, sidelobe level, unambiguous region, CRLB. The conclusion: OTFS simultaneously wins on velocity resolution and matches OFDM on range resolution, without compromising data throughput. OFDM requires dedicated pulse-Doppler processing to match OTFS.

OTFS vs OFDM Radar: Full Comparison

MetricOTFSOFDM (single sym)OFDM (pulse-Doppler)
Range resolution Ξ”R\Delta Rc/(2W)c/(2W)c/(2W)c/(2W)c/(2W)c/(2W)
Velocity resolution Ξ”v\Delta vc/(2Tf0)c/(2 T f_0)∞\inftyc/(2Tdwellf0)c/(2 T_{\text{dwell}} f_0)
Ambiguity shapeThumbtackRidge (in Ξ½\nu)Thumbtack (after NPRIN_{\text{PRI}} pulses)
Peak sidelobe (unwindowed)βˆ’13-13 dB (localized)βˆ’13-13 dB (along ridge)βˆ’13-13 dB (localized)
Data during sensing?Yes (ISAC)Yes (ISAC)No (dedicated radar mode)
Preserves data rate?YesYes (per symbol)No
CRLB range accuracyc/(2W2ρ)c/(2W\sqrt{2\rho})SameSame
CRLB velocity accuracyc/(2Tf02ρ)c/(2Tf_0\sqrt{2\rho})∞\inftySame as OTFS at equal dwell
Detection algorithm2D DFT on DD grid1D DFT per symbol2D DFT over time Γ— frequency
Recommended use6G ISAC, automotiveRange-only OFDMDedicated radar

Theorem: Quantitative Resolution Gap Between OTFS and OFDM-single-Symbol

At equal time-bandwidth product TWTW, OTFS and OFDM-pulse-Doppler achieve the same range and velocity resolution. However:

  • Per-OFDM-symbol OFDM radar has velocity resolution Ξ”vOFDM=c/(2Tsymf0)\Delta v_{\text{OFDM}} = c/(2 T_{\text{sym}} f_0) where Tsym=1/Ξ”fT_{\text{sym}} = 1/\Delta f. For typical 5G NR (Tsym=33ΞΌT_{\text{sym}} = 33\mus), Ξ”vβ‰ˆ900\Delta v \approx 900 m/s β€” useless for any practical radar.
  • OTFS achieves Ξ”v=c/(2Tf0)\Delta v = c/(2 T f_0) where T=NTsymT = N T_{\text{sym}} is the full frame duration. Factor NN better.
  • OFDM pulse-Doppler with NPRI=NN_{\text{PRI}} = N pulses achieves Ξ”vPRI=c/(2NTPRIf0)\Delta v_{\text{PRI}} = c/(2 N T_{\text{PRI}} f_0). Matches OTFS if TPRI=TsymT_{\text{PRI}} = T_{\text{sym}} and NPRI=NN_{\text{PRI}} = N β€” but this is dedicated radar mode with no data.

The structural advantage: OTFS integrates the NN-symbol time extension into the waveform itself; OFDM pulse-Doppler requires switching to a dedicated radar mode. ISAC is the unification of data+sensing that OTFS offers natively.

OFDM per-symbol thinks "range only" because its ambiguity is a ridge along Doppler. OFDM-pulse-Doppler thinks "range and Doppler, in separate modes": first send data, then send radar pulses. OTFS thinks "data and Doppler, simultaneously": the data grid is the Doppler resolution cell.

Quantitatively: given a fixed communications resource budget of (W,T)(W, T), OTFS hits the information-theoretic resolution limits for both range and Doppler; OFDM hits them only for range, and would need to sacrifice data for Doppler.

Key Takeaway

OTFS is the natural ISAC waveform. For the same (W,T)(W, T) budget, OTFS matches OFDM-radar (dedicated pulse-Doppler mode) while simultaneously carrying data. OFDM can do ISAC only by compromising β€” either low-resolution velocity (per-symbol) or switching to radar-only mode (pulse-Doppler). OTFS has no such compromise: one waveform, both data and sensing, at the information-theoretic limits.

Side-by-Side: OTFS vs OFDM Ambiguity Shapes

Plot the ambiguity functions ∣As(Ο„,Ξ½)∣2|A_s(\tau, \nu)|^2 of OTFS and OFDM (single symbol) as 2D heatmaps. OTFS: thumbtack at origin. OFDM: narrow ridge in Ο„\tau, broad plateau in Ξ½\nu. Change parameters to see how the shapes respond.

Parameters
64
16

Example: Automotive ISAC Link Budget

An automotive ISAC system at f0=77f_0 = 77 GHz with W=100W = 100 MHz, T=3T = 3 ms, needs to detect a pedestrian at range 50 m with velocity 1.4 m/s (walking speed) and a truck at 100 m with velocity 25 m/s. Compute the OTFS resolutions and confirm detectability.

⚠️Engineering Note

Three Operational Advantages Over OFDM-Radar

  1. Single coherent observation: OTFS observes the scene for the full frame duration TT; OFDM observes it per-symbol (short snapshot). OTFS gets NN-fold better Doppler resolution at the same data rate.

  2. No mode switching: OTFS's ISAC is native β€” the transmit waveform is both data-bearing and radar-compatible. OFDM requires a scheduler to allocate some time for data, some for radar pulses. Mode switching introduces latency and fragmentation, detrimental for real-time applications.

  3. Thumbtack ambiguity: even if OFDM used dedicated pulse- Doppler, its per-symbol ambiguity is a ridge β€” good range, poor velocity. OTFS's thumbtack at the full-frame level gives clean range-velocity detection without the ridge-sidelobe masking issue.

Consequence: OTFS ISAC achieves the radar performance of dedicated pulse-Doppler mode while running continuously at the data rate of OFDM. This is the basis for the 6G ISAC proposal.

Practical Constraints
  • β€’

    OTFS integrates sensing at 0 data rate penalty

  • β€’

    OFDM + pulse-Doppler: sacrificed data during radar mode

  • β€’

    OTFS thumbtack: simultaneously good range and velocity

OTFS-ISAC Use Cases Compared

Application(W,T,f0)(W, T, f_0)(Ξ”R,Ξ”v)(\Delta R, \Delta v)OTFS advantage
Automotive CRUISE(100 MHz, 3 ms, 77 GHz)(1.5 m, 0.65 m/s)Continuous ISAC, no radar mode
Gesture recognition(500 MHz, 10 ms, 60 GHz)(30 cm, 0.25 m/s)Sub-second response
Indoor positioning(50 MHz, 4 ms, 5 GHz)(3 m, 7.5 m/s)IoT-device tracking
UAV surveillance(200 MHz, 10 ms, 28 GHz)(75 cm, 0.54 m/s)Simultaneous comms + detection
Healthcare monitoring(2 GHz, 5 ms, 60 GHz)(7.5 cm, 0.5 m/s)Heart rate via Doppler