PaperlayerView exact revision on arXiv

Source-linked AI summary

Toward Millimeter Wave Joint Radar-Communications: A Signal Processing Perspective

Kumar Vijay Mishra, Bhavani Shankar M. R., Visa Koivunen, Björn Ottersten, Sergiy A. Vorobyov

arXiv:1905.00690v2eess.SP

TL;DR

The paper addresses inefficient spectrum use and high RF-chain power consumption in joint radar-communications systems. It develops waveform, multiplexing, and multi-objective design approaches, reporting simultaneous cm-level range resolution and Gbps data rates alongside recovery of JRC parameters.

  • Problem

    Fragmented spectrum allocation and unacceptably high power consumption motivate joint radar-communications designs that share spectral and hardware resources.

  • Method

    The paper examines Pareto-optimal design weights, DPSK-based waveforms, multiplexing strategies, and joint coding for radar-communications systems.

  • Results

    cm-level range resolution and a Gbps data rate are achieved simultaneously, while matched-filtered waveforms remain mathematically identical and enable recovery of all JRC parameters.

  • Takeaways & Limitations

    Joint coding can increase radar pulse integration time for enhanced velocity estimation and support detailed detection of automotive targets and pedestrian body movements.

Abstract

from arXiv · show

Synergistic design of communications and radar systems with common spectral and hardware resources is heralding a new era of efficiently utilizing a limited radio-frequency spectrum. Such a joint radar-communications (JRC) model has advantages of low-cost, compact size, less power consumption, spectrum sharing, improved performance, and safety due to enhanced information sharing. Today, millimeter-wave (mm-wave) communications have emerged as the preferred technology for short distance wireless links because they provide transmission bandwidth that is several gigahertz wide. This band is also promising for short-range radar applications, which benefit from the high-range resolution arising from large transmit signal bandwidths. Signal processing techniques are critical in implementation of mmWave JRC systems. Major challenges are joint waveform design and performance criteria that would optimally trade-off between communications and radar functionalities. Novel multiple-input-multiple-output (MIMO) signal processing techniques are required because mmWave JRC systems employ large antenna arrays. There are opportunities to exploit recent advances in cognition, compressed sensing, and machine learning to reduce required resources and dynamically allocate them with low overheads. This article provides a signal processing perspective of mmWave JRC systems with an emphasis on waveform design.

I. INTRODUCTION

mmWave JRC addresses scarce and fragmented spectrum by jointly supporting radar and communications, whose wider-band operation improves data rates and sensing resolution. The mmWave band offers wide bandwidth but imposes strong attenuation, hardware, processing, and channel-adaptation constraints.

  • Motivation: Spectrum fragmentation and rising demand motivate radar-communications sharing, since both applications benefit from wider spectrum.Existing systems also create mutual interference across allocated bands.
  • mmWave Opportunity: mmWave communications provide several GHz of bandwidth for near-field links, while wide bandwidth gives mmWave radar high range resolution.The band supports high-data-rate applications and short-range sensing.
  • Channel Constraints: mmWave signals suffer strong attenuation and lower diffraction, limiting communications coverage and making mmWave radar primarily short-range.The same propagation environment reduces multipath severity.
  • Implementation Constraints: Wide bandwidth requires expensive high-rate ADCs and low-complexity processing, while sparse time and angular channels create opportunities for compressed sensing.Narrowband assumptions may fail when bandwidth is broad relative to carrier frequency.
  • Implementation Constraints: Large arrays and narrow antenna spacing make one RF-IF chain per antenna infeasible, favoring analog or hybrid beamforming.The relevant element spacing is at most λ/2 and can be 0.5–5 mm for mmWave carriers.
  • Channel Constraints: Short coherence times and blockage make beam realignment, feedback, and waveform adaptation difficult in dynamic mmWave JRC environments.The paper notes typical coherence times of nanoseconds and broad Doppler requirements for radar.

A. Communications Channel

The paper models mmWave communications and radar channels using path delays, Doppler shifts, gains, fading, and scattering centers. Radar models can represent multiple reflections from extended targets, while clustered and correlated scenarios remain insufficiently examined.

  • Communications Channel: The LOS communications channel is parameterized by path loss, delay, Doppler shift, and path gain, with γ≈2 for several mmWave LOS scenarios.The free-space model incorporates transmitter and receiver antenna gains and wavelength-dependent attenuation.
  • Radar Channel: The doubly selective mmWave radar channel captures time- and frequency-selective propagation and can include Rician fading.The model also covers spiky special cases and Swerling III/IV target approximations.
  • Radar Channel: Each virtual radar scattering center is described by distance, delay, velocity, Doppler shift, large-scale gain, and small-scale fading gain.The Doppler shift is defined as ν_l,k=2v_l,k/λ, and large-scale gain depends on radar cross section.
  • Modeling Scope: Clustered channel models can incorporate correlations and extended targets, but these scenarios remain unexamined in detail.This limits the scope of the currently detailed modeling treatment.
  • Radar Channel: Extended-target models represent received signals as multiple reflections from different object parts and may include correlated radar cross sections.This is more appropriate for close-range, high-resolution mmWave radar applications than a single-scatterer model.

C. Channel-Sharing Topologies

mmWave JRC systems range from separate radar and communications entities sharing spectrum to co-designed systems that share hardware, antennas, and waveforms. Their designs manage interference using adaptable transmission and reception resources.

  • Spectral Coexistence: Spectral coexistence keeps radar and communications separate while adapting transmit parameters and mitigating interference for the other system.Limited information exchange may support spectral cooperation without major hardware or standardization changes.
  • Co-design: Spectral co-design uses a single unit for radar and communications while accessing spectrum opportunistically.Fully adaptive software-defined platforms seek to reduce circuitry and increase flexibility.
  • Channel-Sharing Topologies: The topology options include independent systems, shared receivers, transmit-shared systems with a common JRC waveform, and bi-static systems sharing transmitter, receiver, and waveform.The figure also identifies an in-band full-duplex variant using different waveforms with common transmitter and receiver.
  • Interference Management: Interference management senses shared-spectrum conditions and adjusts transmitter and receiver parameters to reduce interference and improve system performance.Available degrees of freedom include antennas, frequency, coding, slots, power, and polarization.
  • Performance Criteria: Communications performance can be characterized by achievable or effective spectral efficiency, with effective rate equal to bandwidth multiplied by effective spectral efficiency.The effective quantity depends on the implemented receiver.

B. Radar Performance Criteria

Radar and communications require task-specific criteria for evaluating JRC designs, while interference mitigation uses spatial processing, beamforming, cancellation, and spectrum-aware adaptation. These methods depend on channel knowledge and receiver capabilities.

  • Radar Performance Criteria: Radar detection uses correct-detection, mis-detection, and false-alarm probabilities, while estimation commonly uses MSE or variance relative to the CRLB.The CRLB is the lower bound on estimation-error variance for unbiased estimators.
  • Radar Performance Criteria: Range, Doppler, angular resolution, coverage, and the number of simultaneously resolved targets are radar design parameters.The ambiguity function characterizes range and velocity discrimination by correlating a waveform with delayed and Doppler-shifted replicas.
  • Interference Management: mmWave JRC transmitters and receivers can use antennas, frequency, coding, slots, power, and polarization to mitigate or avoid mutual interference.Transmit parameters may be adapted using receiver feedback about channel response and SINR.
  • Receiver Techniques: Covariance-based projection separates signal and interference-plus-noise subspaces to process signals in a practically interference-free subspace.The approach uses the receive-array covariance matrix or its estimate.
  • Receiver Techniques: Receive beamforming steers nulls toward interference while preserving gain toward desired signals when their arrival angles differ.Common formulations include MVDR, LCMV, and diagonal loading.
  • Receiver Techniques: Advanced interference cancellation estimates channel state or spectrum properties before subtracting interference from the received signal.The cancellation process requires channel coherence long enough for feedback or estimates to remain current.
  • Receiver Techniques: Some cancellation techniques require knowledge of coexisting systems’ modulation schemes or apply only to digital modulation methods.SIC decodes and subtracts the strongest signal before repeating on weaker residual signals.

2) Transmitter Techniques:

Transmitter techniques co-design radar and communications through waveform, precoder, decoder, and interference-management optimization. Performance criteria balance communications distortion, data rate, radar estimation or detection quality, and available system information.

  • Precoder and decoder designs steer interference away from desired signals while jointly targeting receiver SINR, communications data rates, and radar detector performance.
  • SSSVSP steers interference into singular-vector subspaces associated with zero or negligible singular values for MIMO radar-communications coexistence.
  • Interference Alignment coordinates multiple transmitters so mutual interference occupies only part of receiver signal space, leaving interference-free space for radar and communications.
  • Waveform optimization exploits spatial, temporal, spectral, and polarization degrees of freedom according to performance criteria, CSI, target-scene knowledge, and interference levels.
  • A weighted log-scale combination of effective communications MMSE and radar CRLB quantifies trade-offs and yields Pareto-optimal solutions across design goals.The criteria use Q as the number of detected targets and apply weights to different objectives.
  • Mutual-information maximization improves radar target characterization capacity but does not maximize probability of detection, and characterization and detection generally require different signals.

B. Radar-Centric Waveform design

Radar-centric mmWave JRC waveforms use wideband FMCW, PMCW, and related designs to obtain range, Doppler, and angular information while embedding communications. PMCW offers lower sidelobes and easier hardware implementation than FMCW.

  • Conventional Continuous Wave and Modulated Waveforms: FMCW transmits chirps whose linearly varying frequency supports centimeter-scale range resolution with bandwidths of a few gigahertz.A 4 GHz chirp achieves a range resolution of 3.75 cm.
  • Conventional Continuous Wave and Modulated Waveforms: PMCW uses binary pseudorandom sequences with desirable auto- and cross-correlation properties for radar-centric JRC.
  • Conventional Continuous Wave and Modulated Waveforms: PMCW has lower sidelobes than FMCW and is easier to implement in hardware.
  • PMCW-JRC Model: A bi-static ULA PMCW-JRC transmitter sends repeated code sequences from multiple antennas, while target reflections arrive at multiple receive antennas over a coherent processing interval.
  • PMCW-JRC Model: PMCW-JRC combines radar and communications waveforms in analog hardware before the RF stage.
  • PMCW-JRC Model: The received PMCW-JRC signal superposes target reflections, with beamforming, Doppler, delay, propagation loss, target RCS, and noise represented in the model.

Y PMCW-JRC

PMCW-JRC and multi-carrier designs combine radar sensing with communications through time, spectral, code, and spatial degrees of freedom. These approaches support adaptive processing but face hardware, synchronization, and waveform-envelope constraints.

  • PMCW-JRC: Communications symbols and Doppler parameters are coupled in PMCW-JRC, motivating multiplexing strategies that uniquely identify received-signal parameters.
  • PMCW-JRC: PMCW-JRC time-division multiplexes radar-only and joint radar-communications frames across fractions µ and (1 −µ) of the coherent processing interval.The allocation µ depends on prior knowledge about the target scene.
  • PMCW-JRC: Residual signals after communications-symbol extraction can improve radar target estimates through low-complexity JRC super-resolution algorithms.
  • Multi-Carrier Waveforms: Multi-carrier radar provides spectral and other degrees of freedom for agile adaptation to radar tasks, target types, and radio-spectrum conditions.
  • Multi-Carrier Waveforms: A multi-carrier waveform’s time-varying envelope increases PAPR or PMEPR and can hinder efficient amplifier use at high transmit powers.
  • Multi-Carrier Waveforms: In mmWave radar, low transmit powers and short surveillance ranges make the PAPR issue less severe; coding or subcarrier allocation can reduce it further.
  • Multi-Carrier Waveforms: MCPC preserves subcarrier orthogonality while exploiting spectral and code-domain degrees of freedom, and becomes OFDM when subcarriers are uncoded.
  • Spatial DoFs and Multiple Waveforms: Spatial designs embed communications in radar beampattern sidelobes, antenna-waveform pairings, or spatial signatures to separate signals and mitigate interference.

C. Communications-Centric Waveform design

Communications-centric mmWave JRC designs primarily use OFDM and OFDMA to obtain efficient communications processing and temporal-spectral allocation while supporting radar sensing. Their radar performance depends on cyclic-prefix and waveform-model choices.

  • OFDMA-JRC: OFDM is popular for mmWave JRC because it offers a stable performance in multipath fading and relatively simple synchronization.
  • OFDMA-JRC: OFDMA differentiates users in both time and frequency, providing degrees of freedom in temporal and spectral domains.
  • OFDMA-JRC: OFDM-JRC supports high dynamic range and efficient FFT-based receiver implementation but requires additional processing for communications and radar-centric properties.
  • OFDMA-JRC: The OFDM cyclic prefix simplifies communications equalization by converting a frequency-selective channel into multiple frequency-flat channels.
  • OFDMA-JRC: The cyclic prefix may adversely affect radar range-ambiguity resolution, and radar applications require a duration at least as long as maximum signal travel time.
  • OFDMA-JRC: OFDMA-JRC transmits OFDM symbols across multiple antennas and subcarriers, encoding multiplexed communications and radar DPSK symbols.
  • OFDMA-JRC: Its received samples form a radar data cube across spatial, spectral, and temporal domains, with dimensions determined by antennas, subcarriers, and OFDM symbols.

Y OFDMA-JRC

OFDMA-JRC couples communications symbols with radar range, so frequency-division multiplexing reserves known subcarriers for radar processing. Although PMCW- and OFDMA-JRC receive models become mathematically identical after matched filtering, their waveform-specific properties produce different trade-offs.

  • Y OFDMA-JRC: OFDMA-JRC couples communications symbols with radar range, motivating reserved subcarriers for range estimation.A fraction of OFDMA subcarriers is assigned to radar with known symbols, while the remainder serves JRC.
  • Y OFDMA-JRC: Frequency-division multiplexing allocates µ% of OFDMA subcarriers to radar and the rest to JRC.The remaining OFDMA-JRC receive processing is similar to PMCW-JRC.
  • Y OFDMA-JRC: After matched filtering, both waveforms have mathematically identical receive models and recover JRC parameters using similar super-resolution algorithms.Their degrees of freedom and design spaces remain different before receive processing.
  • Y OFDMA-JRC: PMCW-JRC inherits low sidelobes from stand-alone PMCW radar, whereas OFDMA-JRC differs in its ambiguity-function behavior at equal bandwidth.The comparison is presented in Fig. 3.
  • Y OFDMA-JRC: OFDMA-JRC is robust to inter-channel interference through waveform orthogonality, while PMCW-JRC is more sensitive to the number of users.In networked vehicles, predefined or stored PMCW sequences require less infrastructure and processing than adaptive OFDMA band allocation.

D. Joint Coding

Joint coding designs reuse communications waveforms and coding structures for radar, balancing range, Doppler, and communications performance. mmWave examples exploit complementary sequences, preambles, variable frame timing, and sparse processing.

  • Joint waveform structures: 802.11ad SCPHY preambles use Golay complementary pairs whose summed autocorrelation has a peak of 2N and zero sidelobes.This property supports channel estimation and target detection.
  • 802.11ad-based JRC: A single SCPHY preamble simultaneously provides cm-level range resolution and Gbps data rate.The waveform uses the communications preamble for radar sensing.
  • Joint coding trade-offs: Reserving multiple fixed-length radar preambles improves velocity estimation but prolongs preamble duration and significantly reduces communications data rate.A joint sparsity-based coding scheme is described as minimizing this trade-off.
  • Joint coding trade-offs: Varying frame lengths places radar pulses non-uniformly within a coherent processing interval, forming a virtual block that increases integration time and enhances velocity estimation.Frequency-domain sub-Nyquist processing can achieve the same effect when the channel is sparse.
  • High-resolution sensing: A Doppler-resilient 802.11ad waveform produces high-resolution range-Doppler profiles that distinguish automotive targets and resolve individual vehicle wheels and pedestrian body parts.The signatures are shown in range-time and Doppler-time domains.

E. Carrier Exploitation

Carrier exploitation allocates subcarriers and power using channel, occupancy, communications-rate, SINR, and radar-detection information. Water-filling and Neyman-Pearson strategies provide distinct radar-oriented allocation examples.

  • Adaptive carrier allocation: Adaptive subcarrier selection and power or PAPR control can exploit spectrum awareness while satisfying communications and radar requirements.Available information includes occupancy, channel gains, desired SINR, and power constraints.
  • Resource-sharing motivation: Radar generally uses the full bandwidth for high resolution, whereas communications assigns resource blocks to users according to channel quality and rate or QoS requirements.This difference motivates shared-spectrum carrier exploitation.
  • Adaptive carrier allocation: Carrier allocation can impose minimum communications rate constraints while respecting the radar’s maximum transmit power PT.The optimization distributes subcarrier powers Pk across the shared spectrum.
  • Power-allocation strategies: Water-filling allocates radar power where attenuation is lowest and interference is low by maximizing mutual information between received data and target or channel response.This is one radar-oriented carrier-exploitation solution.
  • Power-allocation strategies: The Neyman-Pearson strategy accounts for channel gains and communications SINR requirements while maximizing radar performance for target detection.Its detection threshold η is associated with a false-alarm constraint α.

V. COGNITION AND LEARNING IN MMWAVE JRC

Cognitive and learning methods help mmWave JRC systems sense changing radio environments, adapt waveforms and resource use, and coordinate competing radar-communications objectives. Their practicality is constrained by fast channel dynamics and large waveform dimensions.

  • Cognitive systems: Cognitive mmWave JRC systems estimate channels, exchange feedback, map spectrum access across locations and frequencies, and adjust operating parameters accordingly.Spectrum cartography supports awareness of access conditions over space, frequency, and time.
  • Open challenges: MmWave JRC requires continuous channel cognition, but channel coherence times can be nanoseconds, limiting the time available for cognitive actions.The paper identifies this fast channel variation as an implementation constraint.
  • Compressed sensing: Compressed-sensing methods can reduce required samples for cognitive processing and support adaptive transmission in disjoint subbands.Vacant subbands can be used by vehicular communications.
  • Fast waveforms: Cognitive waveform algorithms must redesign waveforms on the fly with low computational complexity, despite fast-time radar waveforms containing tens of thousands of samples.One cited spectrally dense waveform design has no more than quadratic complexity.
  • Machine learning: Machine learning can acquire situational awareness for low-latency configuration, including target classification, waveform recognition, and antenna or RF-chain selection.Reinforcement learning can learn policies for coexisting systems through POMDP and RMAB approaches.
  • Game-theoretic solutions: Game-theoretic models represent spectrum sharing as dynamic interactions with conflicting interests, yielding Nash or Stackelberg equilibria under specified utility functions.The solution space is several GHz wide with much lower maximum transmit power than sub-6 GHz.
Loading 1905.00690v2…