PaperlayerView exact revision on arXiv

Source-linked AI summary

Seventy Years of Radar and Communications: The Road from Separation to Integration

Fan Liu, Le Zheng, Yuanhao Cui, Christos Masouros, Athina P. Petropulu, Hugh Griffiths, Yonina C. Eldar

arXiv:2210.00446v2eess.SP

TL;DR

Radar and communications historically developed separately despite shared electromagnetic foundations, motivating a clearer account of how they can converge in ISAC. The paper systematically reviews their signal-processing principles, technological evolution, and signals-and-systems duality, then organizes ISAC around resulting tradeoffs and open challenges. It concludes that increasing frequencies, bandwidths, and antenna-array sizes have brought R&C toward integration while leaving fundamental tradeoff characterization and networked deployment as open problems.

  • Problem

    Radar and communications evolved as separate fields, while their recent convergence raises the need to understand the signal-processing foundations and tradeoffs of ISAC.

  • Method

    The paper systematically overviews R&C history from a signal-processing perspective, covering their duality, fundamental principles, spectrum and antenna-array evolution, and ISAC frameworks.

  • Results

    The review shows that higher frequencies, wider bandwidths, and expanded antenna arrays have driven R&C from separation toward ISAC, with shared signal-processing structures and tradeoffs.

  • Takeaways & Limitations

    ISAC links R&C through interchangeable signal-and-system viewpoints and resource-sharing tradeoffs, motivating integrated signal-processing research across spectral and spatial domains.

Abstract

from arXiv · show

Radar and communications (R&C) as key utilities of electromagnetic (EM) waves have fundamentally shaped human society and triggered the modern information age. Although R&C have been historically progressing separately, in recent decades they have been converging towards integration, forming integrated sensing and communication (ISAC) systems, giving rise to new, highly desirable capabilities in next-generation wireless networks and future radars. To better understand the essence of ISAC, this paper provides a systematic overview on the historical development of R&C from a signal processing (SP) perspective. We first interpret the duality between R&C as signals and systems, followed by an introduction of their fundamental principles. We then elaborate on the two main trends in their technological evolution, namely, the increase of frequencies and bandwidths, and the expansion of antenna arrays. We then show how the intertwined narratives of R&C evolved into ISAC, and discuss the resultant SP framework. Finally, we overview future research directions in this field.

I. INTRODUCTION

Radar and communications both exploit electromagnetic waves but historically evolved as separate fields under different application constraints. This paper traces their signal-processing development toward ISAC through shared models, spectrum expansion, and larger antenna arrays.

  • I. INTRODUCTION: Electromagnetic waves support both physical-target information acquisition and artificial-information delivery over long distances.Radar extracts quantities such as range, velocity, and angle, while communications deliver texts, voices, images, and videos.
  • I. INTRODUCTION: Radar and communications developed separately despite their common origins because their applications imposed different constraints.Their later convergence is associated with similarities in hardware architecture, channel characteristics, and signal-processing methods.
  • I. INTRODUCTION: Their technological evolution follows two trends: higher frequencies and larger bandwidths, and expansion from single antennas to massive arrays.The paper organizes its historical overview around spectrum engineering and antenna-array technologies before discussing ISAC.
  • I. INTRODUCTION: Radar and communications share a transmitter-channel-receiver system, but radar focuses on target information acquisition while communications focus on information delivery.Radar may use colocated or separated transmitters and receivers, whereas communication transmitters and receivers are usually separated.
  • I. INTRODUCTION: Interchanging signals and systems makes communication and radar signal-processing problems mathematically similar.Communication treats the channel as a transform applied to a transmitted signal, while radar treats target parameters as an input recovered through a waveform-defined system.

B. Linear Gaussian Models

The linear Gaussian model represents received signals as transmitted signals transformed by a channel and corrupted by white Gaussian noise. Radar estimates physical parameters using known waveforms, whereas communications recover encoded information with channel knowledge obtained through pilots.

  • B. Linear Gaussian Models: The general model uses sampled transmit and receive signals, a channel matrix, and white Gaussian noise with variance σ2.The channel can be defined across time-space or time-frequency domains and may depend on physical parameters such as range, angle, and Doppler.
  • B. Linear Gaussian Models: Radar estimates physical target parameters η from received observations using a typically known deterministic transmit signal.Because the radar waveform contains no information, the information-codeword variable can be omitted.
  • B. Linear Gaussian Models: Communications recover codewords ξ from received observations, usually after estimating the channel with reference pilots.The channel may be treated as an unstructured matrix, while physical parameters η need not be the first priority.
  • B. Linear Gaussian Models: The same generic models cover multiple R&C settings, including MIMO radar, narrowband MIMO communications, and OFDM signal models.The paper uses these generic models without fixing a signal domain before introducing more concrete models later.

1) Signal Detection:

Signal detection decides between hypotheses or symbols, while parameter estimation recovers target or channel parameters from received signals. Radar and communications use related estimation structures but different detection criteria and operating knowledge.

  • 1) Signal Detection:: Radar target detection is formulated as binary hypothesis testing between noise-only reception and reception containing target return plus noise.A detector maps the received signal to a real number and compares it with a threshold γ.
  • 1) Signal Detection:: Communication detection estimates information symbols by minimizing error probability, equivalently using a maximum-a-posteriori decision rule.The decision regions depend on the symbols’ prior probabilities.
  • 1) Signal Detection:: Radar Neyman–Pearson thresholds are set by a required false-alarm probability, unlike communication decision regions based on symbol priors.This difference produces distinct detector designs for R&C systems.
  • 2) Parameter Estimation:: Parameter estimation maps received radar signals to target-parameter estimates and evaluates accuracy with mean squared error.For deterministic parameters, unbiased estimates are lower-bounded by the Cramér–Rao bound, while MUSIC and ESPRIT provide lower-complexity practical alternatives.
  • 2) Parameter Estimation:: Communication channel estimation is mathematically similar to radar target estimation because channel entries play the role of unknown parameters.Transmit pilots are known to both transmitter and receiver and support channel recovery.

3) Information Theory:

Information theory provides a common language for communication capacity and radar estimation performance. Through mutual information, MMSE, and differing cooperation assumptions, the paper exposes both shared structure and distinct signal-processing requirements.

  • 3) Information Theory:: Shannon’s channel coding theorem states that rates below the capacity of a discrete memoryless channel are achievable, while rates above it cannot attain arbitrarily small decoding error.Capacity is the maximum mutual information over possible input distributions.
  • 3) Information Theory:: The same input-output framework describes communication recovery of information signals and radar estimation of random target parameters from echoes.In both cases, useful information is recovered from an observed output Y generated from an input X.
  • 3) Information Theory:: Mutual information and MMSE can both be expressed as functions of SNR, linked by the I-MMSE identity for Gaussian channels.The derivative of mutual information with respect to SNR equals half the MMSE for estimating X from Y.
  • 3) Information Theory:: Gaussian inputs maximize mutual information growth under a given SNR, favoring communication but being least favorable for radar estimation.This contrast connects input-distribution choices with different R&C objectives.
  • 3) Information Theory:: Communication uses cooperation between transmitters and receivers, whereas radar sensing is fundamentally uncooperative and often receiver-complexity dominated.Cooperative communications can flexibly distribute processing, while radar transmitters and receivers typically cannot share design complexities.
  • 3) Information Theory:: The paper next examines spectrum and antenna-array evolution to reveal their interplay in spectral and spatial signal processing.These developments provide the technological path toward integrated sensing and communication.

A. Spectrum Characteristics and Management

R&C spectrum evolution trades lower-band propagation advantages for higher-band bandwidth and resolution, while multi-carrier and stepped-frequency methods extend sensing and communication capabilities.

  • The RF EM spectrum spans below 1 MHz to above 100 GHz and supports diverse communication, broadcasting, navigation, and sensing applications.
  • Lower frequency bands support long-range radar surveillance and weather monitoring, while their lower attenuation favors long-distance communications.
  • Higher frequencies increase achievable bandwidth for finer radar range resolution and higher communication data rates, but atmospheric attenuation limits long-range operation.
  • Multi-carrier technologies are widely used in both radar and communications as wideband signaling strategies.
  • In pulsed radar, target range is inferred from echo delay, with range separation generally requiring |τ_l1 − τ_l2| ≥ 1/B_r.The radar transmits pulses with pulse repetition interval T_PRI, bandwidth B_r, and duty cycle τ/T_PRI.
  • Stepped-frequency waveforms sweep carrier frequencies to estimate target distance and speed, while sparse frequency sampling can create sidelobes that motivate compressed-sensing methods.Random frequency selection can synthesize bandwidth DΔf > B; ℓ1-norm minimization and OMP are cited mitigation approaches.

2) Multi-Carrier Communication Signal Processing:

Multi-carrier communication processing represents data across subcarriers and uses Fourier-domain operations for symbol detection, while jointly supporting radar sensing in OFDM-based ISAC.

  • Communication signals occupy bandwidth B_c, with symbol timing set by T_c = 1/B_c and symbols generated using Nyquist-compatible pulse functions.
  • ASK, FSK, and PSK can generate the transmitted communication symbol sequence.
  • Multi-carrier signaling maps symbol sequences across N_c subcarriers using synthesis functions forming a Gabor system, with an N_c-point IDFT generating the signal.
  • The received multi-carrier signal uses delay and Doppler vectors, after which FFT processing supports detection of the individual subcarrier symbols.
  • Time- and frequency-dispersive channels cause ISI and ICI, while pulse-shaping filters reduce these effects but cannot sharply localize energy in both domains because of the Heisenberg Uncertainty Principle.
  • OFDM can serve communication-centric ISAC by extracting data for communications and estimating range-Doppler profiles through 2D-FFT radar processing.

IV. SCALING UP THE ANTENNA ARRAY: THE ROAD FROM SINGLE ANTENNA TO MASSIVE MIMO

R&C antenna arrays evolved from single and phased arrays to digital, massive, hybrid, and distributed architectures, increasing spatial degrees of freedom while introducing cost and deployment trade-offs.

  • Array Structure Evolution: Scaling antenna arrays provides more exploitable propagation-channel degrees of freedom and can improve transmission reliability.
  • Phased Array: Phased arrays use phase shifters connected to a shared RF chain to generate highly directive beams.
  • MIMO (Digital) Array: Digital MIMO arrays use multiple RF chains and can transmit independent signals for directional steering, omni-directional searching, equalization, or multipath exploitation.
  • Massive MIMO Array: Massive MIMO arrays can exhibit channel hardening and favorable propagation, improving communication reliability through a nearly deterministic channel and simplifying signal processing.
  • Hybrid Array: Hybrid analog-digital arrays reduce hardware cost by connecting fewer RF chains to many antennas, balancing phased-array and fully digital MIMO capabilities.
  • Distributed Array: Distributed antennas spread elements across a large area and connect them to a central processor, increasing coverage probability and diversity gain.
  • Distributed Array: Distributed-array channel models depend on target or scatterer coordinates rather than only arrival angles, with geometry determined by system deployment.

B. Signal Processing for MIMO Radar and Communications

MIMO radar uses spatial processing to balance directional gain, waveform diversity, low interceptability, and target-localization resolution across colocated and distributed configurations.

  • Colocated MIMO radar distributes transmitted energy spatially with low-probability-of-intercept waveforms, avoiding narrow-beam scanning at the cost of processing gain.
  • Widely separated antennas exploit spatial variation in target radar cross section, often using non-coherent combining when phase coherency is difficult.
  • Distributed MIMO radar exploits spatial dimensions and rich backscatter to overcome bandwidth limitations and support high-resolution target localization.

3) MIMO Communications:

MIMO communications gain diversity and multiplexing from expanded antenna arrays, while MIMO radar uses waveform diversity; their signals-and-systems duality exposes shared tradeoffs relevant to ISAC.

  • Multiplexing vs. Diversity: MIMO arrays provide diversity gains against deep fading and multiplexing gains through independent subchannels.Diversity uses different Tx–Rx propagation paths, whereas multiplexing sends distinct data streams over independent subchannels.
  • Multiplexing vs. Diversity: NtNr and min {Nt, Nr} are the maximum diversity and multiplexing gains, respectively, for an i.i.d. Rayleigh MIMO channel.The diversity–multiplexing tradeoff is fundamentally a tradeoff between reliability and efficiency.
  • Multiplexing vs. Diversity: Colocated MIMO radar uses waveform diversity across antennas to improve target-parameter identifiability relative to phased arrays.Waveform diversity can be implemented in baseband or RF through phase or frequency coding.
  • Multiplexing vs. Diversity: MIMO communications transmit multiple data-stream signals through spatial channels, whereas MIMO radar sends diverse waveforms through target channels viewed as signals.This interchangeable signals-and-systems view reveals connections and tradeoffs between radar and communications.
  • Statistical vs. Geometrical Channel Representations: Beam training and tracking in mmWave and THz communications can be viewed analogously to target searching and tracking on hybrid-array RF platforms.This analogy supports merging radar and communications into a single ISAC system.
  • ISAC Evolution: ISAC evolution progresses from spectrum sharing, to common hardware, to unified signaling and processing, and ultimately to shared perceptive networking.The levels increasingly integrate R&C resources and infrastructure.

B. ISAC Signal Processing

ISAC signal processing uses one unified signal for communication and sensing, coordinating both functions through a shared transmit-side framework and balancing their performance objectives.

  • ISAC Signal Processing: A unified ISAC signal S serves both R&C, with radar estimating target parameters η and communications recovering S.The radar receiver knows the reference waveform S, while the communication receiver does not.
  • ISAC Signal Processing: The generic ISAC framework jointly forms a baseband signal, up-converts it, propagates it through R&C channels, and preprocesses received data before separate pipelines.Preprocessing includes synchronization, separation, filtering, and transformation.
  • Performance Tradeoff: R&C performance can be represented by a Pareto frontier using metrics such as radar CRB and communication rate.Communication-optimal and radar-optimal corner points define communication-centric and radar-centric extremes.
  • Communication-Centric Design: Communication-centric design applies radar sensing to an existing communication waveform, prioritizing communications.OFDM-based ISAC directly reuses the OFDM communication waveform for both tasks.
  • Communication-Centric Design: Random communication data degrade radar sensing in OFDM-based ISAC and can be mitigated through element-wise division before a 2D-FFT produces the delay-Doppler profile.The division compensates for the data symbols’ effect in the sensing processing chain.
  • Radar-Centric Design: Radar-centric design embeds communication data into radar waveforms while aiming to avoid undue degradation of sensing performance.More recent MAJoRCom signaling randomly allocates carrier frequencies across antennas while preserving orthogonality.

3) Joint Design:

Joint-design ISAC signaling uses optimization to navigate the radar–communication Pareto frontier, balancing communication rate against radar sensing accuracy. Its tradeoffs arise from both signal randomness and the degree of overlap between communication and sensing subspaces.

  • 3) Joint Design:: Joint design formulates ISAC signaling as a scalable optimization problem that approaches arbitrary points on the radar–communication Pareto frontier.The formulation minimizes angle CRB under a communication sum-rate constraint and transmit-energy constraint.
  • 3) Joint Design:: Increasing the communication sum-rate threshold R0 directs more signal power toward users, leaving less power for sensing and producing a higher CRB.The resulting Pareto frontier is obtained by increasing R0.
  • C. Interplay between R&C: The two fundamental ISAC tradeoffs are deterministic versus random signaling and allocation across communication and sensing subspaces.Communications favor random signals for information transmission, whereas radar favors deterministic signals for stable sensing.
  • 1) Deterministic vs. Random Tradeoff:: Embedding more random data increases communication rate but deteriorates radar sensing performance in communication- and radar-centric signaling schemes.The tradeoff reflects the additional degrees of freedom used to carry information.
  • 2) Subspace Tradeoff:: When sensing and communication subspaces overlap, shared resource allocation improves efficiency; orthogonal subspaces prevent resource reuse and provide zero performance gain.Coupling is categorized as weak, moderate, or strong according to the degree of subspace overlap.
  • 2) Subspace Tradeoff:: Varying the user SINR constraint rotates the ISAC signal between communication and sensing subspaces, forming a scalable tradeoff between achievable rate and radar CRB.The numerical example considers a single-target, single-user scenario.
  • 2) Subspace Tradeoff:: Increasing the correlation coefficient ρ from 0 to 1 improves tradeoff performance because greater subspace alignment enables more resource sharing.At ρ = 1, both radar and communication performance reach their optima without jeopardizing one another.

VI. OPEN CHALLENGES AND FUTURE RESEARCH DIRECTIONS

ISAC research still faces open challenges in fundamental tradeoffs, practical signal processing, and networked operation. The paper identifies these boundaries while concluding that ISAC may substantially affect modern society.

  • Fundamental ISAC performance tradeoff: Current ISAC tradeoff results describe only two corner points, leaving the exact Pareto frontier and optimal signaling strategies unresolved.More practical multi-user, multi-target regimes remain at an early research stage.
  • Practical ISAC signal processing: Most ISAC signaling schemes assume ideal conditions, but limited bandwidth, high PAPR, and hardware imperfections constrain practical radar-sensing performance.Examples include 5G NR waveforms, quantized phase shifters, and uncalibrated antenna arrays.
  • Networked ISAC: Networked ISAC requires signal-processing solutions for commercial infrastructures, including clock-level synchronization and full-duplex base-station operation.These requirements support accurate sensing and short-range target detection involving humans and vehicles.
  • Conclusion and future directions: The paper surveys R&C evolution from an SP viewpoint, emphasizing frequency and bandwidth growth, antenna-array expansion, ISAC progress, and major open challenges.Its conclusion frames ISAC as the marriage of radar and communications with potentially large societal impact.
Loading 2210.00446v2…