Source-linked AI summary
Radar and Communication Co-existence: an Overview
Le Zheng, Marco Lops, Yonina C. Eldar, Xiaodong Wang
TL;DR
Radar–communication coexistence is needed because wireless services and sensing applications require overlapping spectrum. This paper surveys signal models, waveform and receiver techniques, and organizes prior work into three coexistence architectures, while identifying deployment and modeling boundaries and future hardware-testing needs.
Problem
Overlapping spectrum demands coexistence mechanisms that support high-rate wireless communication and reliable radar sensing.
Method
The paper reviews existing coexistence results, develops a three-category taxonomy, and discusses signal models, waveform designs, and signal-processing techniques.
Results
The review groups coexistence strategies into spectral overlap, cognition-based disjoint-band allocation, and functional coexistence with one active transmitter.
Takeaways & Limitations
The surveyed approaches provide different ways to share spectrum, including interference mitigation, cognitive avoidance, and combined radar–communication transmission.
Takeaways & Limitations
Existing results include simplified channel models with single or radiation-absorbing objects, while range-spread objects are not accounted for in the open literature.
Abstract
from arXiv · showhide
Increased amounts of bandwidth are required to guarantee both high-quality/high-rate wireless services (4G and 5G) and reliable sensing capabilities such as automotive radar, air traffic control, earth geophysical monitoring and security applications. Therefore, co-existence between radar and communication systems using overlapping bandwidths has been a primary investigation field in recent years. Various signal processing techniques such as interference mitigation, pre-coding or spatial separation, and waveform design allow both radar and communications to share the spectrum. This article reviews recent work on co-existence between radar and communication systems, including signal models, waveform design and signal processing techniques. Our goal is to survey contributions in this area in order to provide a primary starting point for new researchers interested in these problems.
I. INTRODUCTION
The paper reviews radar–communication coexistence and organizes proposed approaches into three architectures: spectral overlap, cognition-based coexistence, and functional coexistence.
- I. INTRODUCTION: The paper reviews existing radar–communication coexistence results and defines a taxonomy of three major architectures.The categories are spectral overlap, cognition-based coexistence, and functional coexistence.
- I. INTRODUCTION: Spectral-overlap systems use active radar and communication transmitters in the same band and must mitigate mutual interference while preserving both functions.The systems may be uncoordinated or coordinated through joint transmission-policy and receiver-strategy design.
- I. INTRODUCTION: Co-design jointly optimizes sensing waveforms and communication codebooks so both radar and communication performance remain satisfactory.This holistic approach allows the systems to negotiate transmit policies and adjust detection or demodulation strategies.
- I. INTRODUCTION: Knowledge-based coexistence schemes presuppose information exchange, a shared fusion center, and access to a common database of channel parameters.These requirements constrain deployment architectures for such schemes.
- I. INTRODUCTION: Cognition-based coexistence learns channel state and avoids spectral overlap by assigning disjoint sub-bands to radar and communication systems.The approach is associated with channel sensing and interference-free corresponding channels.
- I. INTRODUCTION: Functional coexistence uses one active transmitter, combining radar and communication functions in shared hardware without producing interference between separate transmitters.Dual Function Radar Communication systems embed information in radar signals and use beamforming and waveform diversity.
A. System Model
The system model describes a MIMO radar and MIMO communication system sharing frequency resources, with radar codes, communication waveforms, interference channels, targets, clutter, and noise represented explicitly.
- A. System Model: The general scenario pairs a MIMO radar with M_T transmit and M_R receive antennas and a MIMO communication system with N_T transmit and N_R receive antennas.The radar and communication systems may use co-located or non-co-located antennas.
- A. System Model: The MIMO radar transmits M_T coded signals, with each transmit element characterized by a fast-time code of length P_r.Radar transmissions may use coded pulses and slow-time amplitude modulation.
- A. System Model: Nyquist waveforms use bandwidth B = 1/T_r and satisfy R_ψ(kT_r) = δ(k), while full spectral overlap corresponds to L = 1.The effective duration of coded pulses can make their time–bandwidth product much larger than one.
- A. System Model: Practical Nyquist waveforms are generated by truncating ideal strictly band-limited signals, which can introduce aliasing.Allowing excess bandwidth can keep this effect under control.
- A. System Model: Radar waveform choices include single coded pulses, amplitude-modulated pulse trains, multi-antenna sophisticated signals, and multi-antenna pulse trains.The associated transmitter degrees of freedom include fast-time code matrices C and slow-time code matrices G.
- A. System Model: Pulse repetition time T creates range ambiguities because scatterers with delays differing by integer multiples of T contribute to the same range cell.This ambiguity is distinct from the fast-time waveform representation.
- A. System Model: Radar receiver signals contain target returns, communication-to-radar interference, clutter, and noise, whereas communication receiver signals contain communication channels, radar-to-communication interference, and noise.The model distinguishes coordinated settings, where some quantities are known, from uncoordinated settings, where both interference channels and radar signals may be unknown.
B. Uncoordinated design: radar centric
The radar-centric uncoordinated design treats radar as primary and designs its code to maximize radar SINR while constraining interference delivered to coexisting communication users.
- B. Uncoordinated design: radar centric: The radar-centric approach allows communication users to transmit in partial spectral overlap while limiting interference produced by the radar on shared bandwidths.The radar is assumed to use a single coded pulse, and the design focuses on radar performance.
- B. Uncoordinated design: radar centric: Radar processing samples received signals at the Nyquist rate, applies matched filtering with transmitted codes, performs beamforming, and detects Doppler.These stages form the conventional collocated MIMO radar receive chain.
- B. Uncoordinated design: radar centric: Communication users occupy specified frequency bands, and the interference energy delivered to user k is represented by c^H R_k c.The matrices R_k encode the radar interference contribution over the users’ bands.
- B. Uncoordinated design: radar centric: The optimization formulation assumes exogenous interference covariance is known or perfectly estimated.This assumption applies to the signal-independent component of the aggregate interference.
- B. Uncoordinated design: radar centric: The radar code is designed to maximize receive SINR subject to communication-interference, radar-energy, and similarity constraints.The similarity constraint keeps the code close to a reference waveform with prescribed correlation properties.
- B. Uncoordinated design: radar centric: With clutter covariance of rank greater than one, constrained SINR maximization becomes a fractional non-convex problem.The simpler formulation applies when clutter is absent or modeled as a rank-one specular reflection.
C. Uncoordinated design: communication centric
In uncoordinated spectral overlap, communication receivers must jointly estimate radar interference and demodulate data without relying on radar transmit-policy changes. The reviewed approaches model unknown radar delays and couplings, exploit sparsity, and use iterative or sparse-recovery methods to mitigate interference.
- Uncoordinated design: communication centric: Communication systems must provide QoS when radar transmitters do not modify their transmission policies.The motivation is that radar deployment changes may impose major costs on governmental and military agencies.
- Uncoordinated design: communication centric: The considered model allows multiple potentially active radars in full spectral overlap, without prior knowledge of their number, distances, or channel gains.Communication signals may use precoding spanning CDMA and OFDM models, while the discussion specializes to direct constellation-point transmission.
- Uncoordinated design: communication centric: The signal model assumes a single-antenna communication receiver, single-tap channels, a nonsaturating radar front end, and guaranteed frame synchronization.The radar transmitter is typically high-power, but the receiver is assumed not to saturate.
- Uncoordinated design: communication centric: The receiver jointly estimates radar interference parameters and demodulates communication data because these tasks are inherently coupled.The parameters include radar delays and complex coupling coefficients used to subtract the estimated interference.
- Uncoordinated design: communication centric: Iterative joint demodulation and interference estimation can start from direct demodulation and then estimate radar delays and couplings for interference removal.The residual after demodulation contains communication errors, residual radar interference, and noise.
- Uncoordinated design: communication centric: Atomic-Norm minimization exploits two sparsity structures to recover radar components and communication coefficients, with Figure 2 comparing AN-based and CS-based SER results.The radar mixture contains at most MT components with MT ≪ Pr, while the coefficient vector is encouraged to have small ℓ0 support.
- Uncoordinated design: communication centric: A pulse-train radar creates intermittent, high-PAPR interference, and phase-only radar coding can make the interference approximately constant-envelope for narrow-band communication.The statistical model treats interference phase as uniform and radar-interference power as known during affected intervals.
D. Coordinated design
Coordinated design treats radar and communication systems as cooperating components that jointly choose waveforms and communication codebooks, including under full spectral overlap. The section develops null-space and optimization-based strategies to protect communication quality while limiting radar-performance loss.
- Co-design: Co-design coordinates transmit policies, detection or demodulation strategies, radar waveforms, and communication codebooks across the two systems.The approach assumes information exchange between the active radar and communication systems.
- System model: The coordinated model represents full-overlap MIMO radar and communication systems with radar space-time code matrix D and communication signal matrix V.The radar target and clutter responses, interference channels, communication channel, and receiver noises appear in the signal models.
- Spatial coordination: Null-space projection designs radar waveforms orthogonal to selected communication channels, using the MIMO transmitter degrees of freedom to suppress interference.Projection onto a receiver channel’s null space can guarantee zero interference there, but introduces correlation among radar transmit signals and may reduce direction-estimation performance.
- Spatial coordination: Selecting the best interference channel makes the null-space-projected waveform’s target-direction RMSE closer to that of the original radar waveform.The comparison uses the UBest and UWorst channel selections shown through the corresponding HBest and HWorst labels.
- Spatial coordination: A MIMO radar can avoid interference at every communication receiver when its transmit-antenna count exceeds the sum of all receivers’ requested degrees of freedom.Cooperation among base stations and the radar supports interference-free operation under this degrees-of-freedom condition.
- Optimization-based design: Optimization-based coordinated design jointly selects radar code matrices, communication covariance matrices, and radar receive filters under radar and communication QoS constraints.The resulting optimization is typically non-convex, and the radar and communication figures of merit depend on transmitted signals and channel parameters.
A. Environment sensing techniques
Environment sensing techniques let radar and communication systems learn spectral conditions and adapt transmission or sampling without necessarily relying on fixed coordination. The surveyed approaches include communication-receiver sensing, radar-receiver sensing, and sub-Nyquist cognitive radar architectures.
- A. Environment sensing techniques: Knowledge-based spectrum sharing uses environment sensing to determine transmit policies, while newer approaches can learn without pilots or coordination.The surveyed scenarios place sensing at either the communication receiver or the radar receiver.
- A. Environment sensing techniques: A sub-Nyquist cognitive radio can detect sparse communication signals at very low rates using a modulated wideband converter.The MWC identifies occupied bands for subsequent radar adaptation.
- A. Environment sensing techniques: The SpeCX prototype combines an MWC-based cognitive radio receiver, a communication receiver, and a cognitive radar to reconstruct spectra and display coexisting radar operation.Its communication display includes 120 MHz low-rate samples from one MWC channel and sub-Nyquist full-spectrum reconstruction.
- A. Environment sensing techniques: After occupied bands are identified, a cognitive radar transmits several narrowband signals in vacant bands and uses Xampling with compressed beamforming for high-resolution delay and Doppler estimation.The resulting radar occupies small total bandwidth and supports low-rate, low-power receivers.
B. Knowledge-based design
Knowledge-based design uses sensed channel or spectral information to shape radar transmission around communication occupancy and radio-environment constraints. The reviewed methods optimize interference-related or waveform-specific objectives while preserving radar requirements.
- B. Knowledge-based design: Radar waveforms can be constrained to the null space of the interference channel between radar transmitters and communication receivers.This approach aims to eliminate radar interference at the communication receiver through spatial degrees of freedom.
- B. Knowledge-based design: After identifying communication support Fc, the radar selects frequencies outside the union of Fc and the radio environment map Fr.The selection objective is to maximize correct detection probability under fixed false-alarm probability.
- B. Knowledge-based design: Structured sparsity designs the radar frequency set to maximize SINR or minimize spectral power in undesired spectrum, subject to additional transmission requirements.The cited requirements include transmit-energy and range-sidelobe constraints.
- B. Knowledge-based design: Spectral notching minimizes transmit energy in specified frequency bands while maintaining desirable envelope and sidelobe characteristics.Spectrally disjoint waveforms require metrics other than SINR because interference does not drive their design.
A. Embedding data into radar waveforms
Dual-function radar-communication systems combine sensing and communication in one hardware platform by embedding information into radar transmissions. The reviewed embedding strategies manipulate waveform selection, phase, sidelobe amplitude, or multiple orthogonal waveforms while accounting for radar-performance effects.
- A. Embedding data into radar waveforms: DFRC combines radar and communication transmitters in one hardware platform, typically using a shared waveform or transmitter designed to support both functions.This functional co-existence avoids spectrum overlap and resource negotiation because information is embedded in the radar signal.
- A. Embedding data into radar waveforms: Waveform-diversity embedding selects one of K = 2^Nb orthogonal radar waveforms pulse by pulse to transmit Nb information bits per pulse.The selected waveform represents the communication symbol.
- A. Embedding data into radar waveforms: Phase-modulation embedding controls the phase of a constant-envelope vector carrying communication symbols through a single-antenna radar waveform.Radar phase modulation controls range-sidelobe modulation, trading off BER and/or data throughput.
- A. Embedding data into radar waveforms: Sidelobe-amplitude modulation associates each communication symbol with a beamforming weight vector and optimizes the resulting transmit beam pattern.The formulation specifies surveillance and communication sidelobe regions, user-controlled sidelobe limits, and communication-direction sidelobe levels.
- A. Embedding data into radar waveforms: Multi-waveform amplitude-shift keying transmits Nb orthogonal waveforms simultaneously, assigning each waveform to a high- or low-beamforming vector according to one information bit.The total transmit energy is divided equally among the Nb waveforms.
B. Radar employing communication waveforms
Radar employing communication waveforms reuses transmissions from communication systems for sensing, including passive radar and opportunistic radar based on 802.11ad. The surveyed methods address direct-path interference, incomplete waveform knowledge, and target detection or localization under different cooperation assumptions.
- B. Radar employing communication waveforms: 802.11ad opportunistic radar exploits millimeter-wave communication characteristics for short-range obstacle detection, using a co-located receiver with access to timing and transmitted-signal information.The standard uses a preamble of concatenated complementary Golay codes and a data payload.
- B. Radar employing communication waveforms: GLRT variants for 802.11ad sensing differ in prior knowledge, payload access, and whether the target coefficient is modeled as a nuisance parameter.The GLRT also produces maximum-likelihood estimates of unknown parameters as a by-product of composite-hypothesis testing.
- B. Radar employing communication waveforms: Figures 7 and 8 illustrate detection and localization performance for short-range obstacle sensing as functions of target range and SNR per bit.Figure 7 fixes the false-alarm probability at Pfa = 10^-4, while Figure 8 reports ranging accuracy.
- B. Radar employing communication waveforms: The reported channel models assume either a single object or an object shielding further obstacles, and they do not account for range-spread objects.This scope is especially relevant because the range resolution is on the order of decimeters.
- B. Radar employing communication waveforms: Passive radar uses communication, broadcast, or navigation transmissions instead of a dedicated radar transmitter, requiring reference and surveillance channels.The reference channel supplies the direct-path waveform for matched filtering, while surveillance channels collect target reflections.
- B. Radar employing communication waveforms: Passive-radar performance can improve when known communication modulation is exploited to demodulate data symbols before detection and estimation.The cited approach obtains better accuracy than directly using the reference-channel signal.
- B. Radar employing communication waveforms: Passive radar is generally inferior to active radar because its waveforms, spatial beampatterns, and transmit power are non-optimal.Commensal radar addresses this setting by designing communication signals for both information transfer and target localization.
V. CONCLUSIONS
The review groups spectrum-sharing strategies into three categories, compares their advantages and disadvantages, and identifies real-world testing as a future need.
- V. CONCLUSIONS: Three categories organize the reviewed coexistence strategies: spectral overlap, cognitive disjoint sub-band assignment, and functional coexistence with one active transmitter.The latter two approaches avoid mutual interference through spectrum partitioning or transmitter coordination.
- V. CONCLUSIONS: The paper outlines each category’s basic ideas, advantages, disadvantages, and illustrative performance examples.
- V. CONCLUSIONS: Hardware prototypes should be deployed and tested on real data to assess performance under varied noise, clutter, and interference conditions.