Integrated Sensing and Communications: Towards Dual-functional Wireless Networks for 6G and Beyond

Fan Liu; Yuanhao Cui; Christos Masouros; Jie Xu; Tony Xiao Han; Yonina C. Eldar; Stefano Buzzi

Source-linked AI summary

Integrated Sensing and Communications: Towards Dual-functional Wireless Networks for 6G and Beyond

Fan Liu, Yuanhao Cui, Christos Masouros, Jie Xu, Tony Xiao Han, Yonina C. Eldar, Stefano Buzzi

arXiv:2108.07165v1Aug 16, 2021eess.SP

TL;DR

ISAC must reconcile the different spatial-resource needs of sensing and communications while existing overviews cover only parts of its theoretical framework. This paper surveys ISAC use cases, tradeoffs, waveform and receiver designs, and perceptive-network integration, concluding that joint designs and mutual assistance can improve dual-functional wireless networks.

Problem
ISAC faces fundamental tradeoffs because communications exploit available spatial degrees of freedom, whereas sensing may be harmed by some propagation paths.
Method
The paper comprehensively surveys ISAC use cases, theory, S&C tradeoffs, waveform and receiver processing, and mutual assistance in perceptive mobile networks.
Results
Joint design spans sensing-optimal to communication-optimal operation, while its fast-time coding and sensing-specific constraints improve data rate and sensing quality over separated designs.
Takeaways & Limitations
The overview positions mutual assistance between communications and sensing, including coordinated downlink bi-static sensing, as a promising direction for future perceptive wireless networks.
Takeaways & Limitations
Most sensing-centric designs use slow-time coding, limiting their application to scenarios requiring low or moderate data rates.

Abstract

from arXiv · show

As the standardization of 5G is being solidified, researchers are speculating what 6G will be. Integrating sensing functionality is emerging as a key feature of the 6G Radio Access Network (RAN), allowing to exploit the dense cell infrastructure of 5G for constructing a perceptive network. In this paper, we provide a comprehensive overview on the background, range of key applications and state-of-the-art approaches of Integrated Sensing and Communications (ISAC). We commence by discussing the interplay between sensing and communications (S&C) from a historical point of view, and then consider multiple facets of ISAC and its performance gains. By introducing both ongoing and potential use cases, we shed light on industrial progress and standardization activities related to ISAC. We analyze a number of performance tradeoffs between S&C, spanning from information theoretical limits, tradeoffs in physical layer performance, to the tradeoff in cross-layer designs. Next, we discuss signal processing aspects of ISAC, namely ISAC waveform design and receive signal processing. As a step further, we provide our vision on the deeper integration between S&C within the framework of perceptive networks, where the two functionalities are expected to mutually assist each other, i.e., communication-assisted sensing and sensing-assisted communications. Finally, we summarize the paper by identifying the potential integration between ISAC and other emerging communication technologies, and their positive impact on the future of wireless networks.

I. INTRODUCTION … 3) Parallel Development of Radar and Communication:

B5G/6G applications require both high-quality connectivity and accurate, robust sensing, motivating ISAC’s unification of information transfer and extraction. ISAC developed from early radar-embedded communications through converging radar and communication technologies, exposing opportunities for shared architectures and mutual performance gains.

A. Background and Motivation: B5G/6G applications such as connected intelligence, smart cities, connected vehicles, and remote healthcare demand high-quality connectivity alongside accurate, robust sensing.Sensing is expected to play an increasingly significant role in future wireless networks.
A. Background and Motivation: ISAC seeks to unify sensing and communication, enabling direct tradeoffs and mutual performance gains while improving spectral and energy efficiency.Sensing extracts information from noisy observations, whereas communication transfers information through tailored signals and noisy reception.
B. Historical View of ISAC: ISAC dates to the 1960s, when information was embedded in radar pulse intervals using pulse interval modulation; later names included RadCom, JCR, and JRC.This early integration demonstrates that joint radar-communication signaling predates its recent rise in academic and industrial attention.
1) The Birth of Radar:: Radar emerged in the first half of the twentieth century for geophysical monitoring, air-traffic control, weather observation, and defense and security surveillance.The US Navy first used the term RADAR in 1939.
1) The Birth of Radar:: Mechanical rotary radars faced limited multifunctionality, flexibility, and jamming resistance, motivating phased arrays that electronically scanned space through spatial control.The first microwave phased-array antenna is attributed to Luis Alvarez, and Germany’s GEMA built the practical FuMG 41/42 Mammut radar.
2) How Radar and Communication Inspire Each Other:: Radar and communications began merging in the 1990s–2000s through ONR’s AMRFC program, which targeted integrated RF front-ends partitioning antennas among radar, communications, and electronic warfare.Subsequent developments included collocated MIMO radar, phased-MIMO radar, and hybrid analog-digital communication architectures.
2) How Radar and Communication Inspire Each Other:: Communication technologies increasingly supported radar sensing: OFDM enabled decoupled delay-Doppler processing, while MIMO concepts contributed degrees-of-freedom and diversity to radar theory.Shared-spectrum research also pursued coexistence between radar and commercial communications, including DARPA’s SSPARC project.
3) Parallel Development of Radar and Communication:: Radar and communication evolved largely in parallel, creating duplicated devices and shared signal-processing structures that motivate deeper ISAC integration and its formalized scope.Parallel examples include phased arrays, MIMO systems, beamforming, detection, estimation, and angle processing; later work also connected massive MIMO and mmWave communications with radar.

4) Convergence of S&C: … 2) Smart Home and In-Cabin Sensing:

The paper frames ISAC as the convergence of sensing and communications through shared transmissions, resources, and infrastructure, and develops this vision from technical foundations to applications and research directions. It highlights perceptive cellular networks, indoor sensing, and spatial-aware computing as representative paths toward future wireless networks.

4) Convergence of S&C:: Radar and communications increasingly converge because both use high-frequency bands, large antenna arrays, similar channel characteristics, and related signal processing.At mmWave frequencies, sparse, Line-of-Sight-dominated channels align more closely with physical geometry and massive MIMO.
C. ISAC: A Paradigm Shift in Wireless Network Design: ISAC integrates sensing and communication to share wireless resources, reduce duplicated transmissions, devices, and infrastructure, and enable mutual benefits.The paper identifies these benefits as Integration Gain and Coordination Gain.
D. Structure of the Paper: The paper presents a comprehensive ISAC overview spanning applications, industrial activities, theoretical limits, S&C performance tradeoffs, waveform design, and receive signal processing.It positions the overview as broader than prior works focused on individual aspects of ISAC.
II. APPLICATIONS AND INDUSTRIAL PROGRESS: The application section extends prior use-case studies with seven potential ISAC scenarios, key use cases, industrial activities, and research efforts connecting academia with industry.These applications support the paper’s vision of future wireless networks.
1) Sensing as a Service:: Dense 5G cellular infrastructure can be reused for sensing with small hardware, signaling, and standards modifications, including reuse of reference or synchronization signals as sensing waveforms.This enables sensing integration into IoT devices and cellular networks rapidly and cheaply.
1) Sensing as a Service:: ISAC-enabled cellular networks can become perceptive networks that improve localization, extract environmental information, provide RF imaging, and monitor or manage non-cooperative and cooperative UAVs.Additional Doppler processing and multipath information improve localization, while cellular sensing can support drone monitoring, control, and swarm navigation.
2) Smart Home and In-Cabin Sensing:: ISAC-enabled IoT supports indoor applications including activity recognition, health care, home security, and driver-attention monitoring using privacy-preserving wireless signals.Wireless amplitude and phase variations can reveal human presence, proximity, falls, sleep, breathing, and daily activities.
2) Smart Home and In-Cabin Sensing:: Spatial-aware computing exploits geometric relationships among dense IoT devices to coordinate household products, analyze movement, understand mobility patterns, expedite handovers, and support augmented virtual reality.Spatial relationships between devices and access points can inform operations beyond SINR-only considerations.

3) Vehicle to Everything (V2X): … 1) Sensing Performance Metrics:

The paper presents ISAC applications spanning connected vehicles, smart factories, remote sensing, environmental monitoring, and touchless human–computer interaction, alongside industry standardization and foundational sensing metrics. It frames ISAC through sensing-task definitions and performance tradeoffs across information theory, physical-layer, propagation, and cross-layer dimensions.

3) Vehicle to Everything (V2X):: ISAC-aided V2X can support vehicle platooning, secure access, SLAM, and extended sensing through RSU networks while reducing antennas, system size, weight, and power consumption.RSUs can serve multiple vehicles simultaneously and provide sensing beyond a passing vehicle’s line of sight and field of view.
3) Vehicle to Everything (V2X):: High latency in multi-hop V2V communications can desynchronize long, dynamic platoons, whereas RSU-based sensing offers a more reliable approach to forming and maintaining them.Conventional platooning mainly uses cooperative adaptive cruise control in a leader-follower framework.
4) Smart Manufacturing and Industrial IoT:: In smart factories, ISAC combines ultra-fast, low-latency communications with sensing for navigation, coordination, environmental mapping, and potentially reduced signaling overhead.The envisioned technology incorporates elements such as swarm navigation, platooning, and imaging.
5) Remote Sensing and Geoscience:: Spaceborne SAR systems use chirp or OFDM waveforms, into which communication data can be embedded to broadcast low-speed streams or provide covert battlefield communications.These radar systems support high-resolution, all-weather, day-and-night imaging for monitoring, change detection, 4D mapping, security, and planetary exploration.
5) Remote Sensing and Geoscience:: Drone or satellite swarms can exchange sensed information and form a large virtual aperture, enabling high-resolution, low-altitude airborne imaging with swarm-based SAR algorithms.Drones also support disaster damage assessment, search and rescue, and emergency communications, but their heavy payloads limit endurance.
6) Environmental Monitoring:: Wireless propagation characteristics can reveal environmental information, with frequency-dependent sensitivity enabling monitoring such as humidity inference from city-wide mmWave path-loss data.High-frequency mmWave signals are sensitive to humidity because they are near water-vapor absorption bands.
7) Human Computer Interaction (HCI):: Wireless sensing can recognize gestures through time, frequency, and Doppler variations, enabling touchless interfaces such as projected virtual keyboards, while micro-Doppler accuracy and temporal resolution remain challenges.Gesture-based touchless interaction is positioned as a potential new HCI application for smartphones and other user equipment.
B. Industry Progress and Standardization: Major companies, IEEE, and 3GPP have advanced ISAC visions and specifications, while ISAC performance analysis covers information-theoretical, PHY, propagation, and cross-layer tradeoffs.IEEE 802.11 established sensing study groups and Task Group IEEE 802.11bf to enhance sensing through 802.11-compliant waveforms.

2) Communication Performance Metrics: · B. Information-Theoretical Limits

Communication performance is framed through efficiency and reliability, while ISAC information theory connects communication metrics with sensing estimation and characterizes fundamental rate–distortion tradeoffs. These results show how resource sharing and feedback-based models support joint communication and sensing objectives.

2) Communication Performance Metrics:: Communication PHY performance is evaluated through efficiency and reliability under limited wireless resources and harmful channel effects.Efficiency concerns successful information delivery; reliability concerns reducing or correcting erroneous information bits.
2) Communication Performance Metrics:: Spectral efficiency and energy efficiency measure achievable rate per unit bandwidth or energy, with units bit/s/Hz or bits/channel use, and bit/s/J.
2) Communication Performance Metrics:: Reliability is commonly measured by outage probability, bit error rate (BER), symbol error rate (SER), and frame error rate.
B. Information-Theoretical Limits: Information theory provides communication-system tools, but sensing lacks equally clear information-theoretical performance definitions, motivating new analytical techniques for ISAC.
B. Information-Theoretical Limits: The mutual information derivative with respect to snr equals half the MMSE regardless of input statistics, linking communication information theory with sensing estimation theory.This relation connects the communication metric mutual information and the sensing metric minimum mean squared error (MMSE).
B. Information-Theoretical Limits: A Gaussian input maximizes mutual information for Gaussian channels and MMSE, favoring communication while being least favorable for sensing.
B. Information-Theoretical Limits: The capacity-distortion framework jointly represents information transmission and channel-state description, defining achievable rate-distortion pairs through decoding reliability and state-estimation error.The framework was introduced for channels with state information and uses a distortion measure between the state and its estimate.
B. Information-Theoretical Limits: Under a state-dependent Gaussian channel, the Pareto-optimal rate-distortion boundary is achieved by splitting transmit power between pure information delivery and scaled channel-state estimation.The power-sharing strategy allocates γP and (1 −γ)P to the two objectives, respectively.

C. Tradeoff in PHY · 1) Tradeoff between Native S&C Metrics:

The PHY layer exposes tradeoffs between native sensing and communication metrics when wireless resources are shared or reused. Detection–rate allocation illustrates power-sharing conflicts, while common-waveform design reveals rank, interference, and dedicated-stream mechanisms that balance estimation and communication.

C. Tradeoff in PHY: PHY tradeoffs can be studied directly through native sensing and communication metrics or through a new sensing information metric matched to communication measures.The paper reviews both approaches to analyzing S&C integration.
1) Tradeoff between Native S&C Metrics:: Sensing performance is typically characterized by detection probability and MSE, with CRB providing an analytical lower bound when closed-form MSE expressions are unavailable.CRB represents the lower bound on the variance of unbiased estimators.
1) Tradeoff between Native S&C Metrics:: Sharing transmit power between radar and communication signals creates a detection-probability versus achievable-rate tradeoff under a communication-rate threshold and total-power constraint.The optimization allocates PR and PC subject to R ≥ Rth and PR + PC = PT.
1) Tradeoff between Native S&C Metrics:: A common waveform can fully reuse temporal, spectral, power, and signaling resources for estimation and multi-user communication, improving efficiency over non-overlapping resource designs.The resulting design estimates the target response matrix while supporting communication utility constraints.
1) Tradeoff between Native S&C Metrics:: Full-rank sensing waveforms require all available degrees of freedom, whereas conventional MU-MISO downlink transmission typically produces rank-deficient waveforms when K ≤ Nt.Rank deficiency makes the sample covariance matrix non-invertible and prevents the required unbiased estimation or MLE.
1) Tradeoff between Native S&C Metrics:: Adding at least Nt − K dedicated sensing streams makes the transmit waveform full rank but introduces interference into users’ received communication data.The per-user SINR denominator includes both multi-user interference and interference from dedicated sensing streams.
1) Tradeoff between Native S&C Metrics:: Dedicated sensing streams can improve MIMO radar beampattern design while guaranteeing communication-user SINR compared with schemes using only communication precoding.The communication waveform also contributes to monostatic sensing rather than interfering with it.

2) Tradeoff between Novel Sensing Metrics and Communication Metrics:

The section introduces sensing-specific metrics that quantify target distinguishability and uncertainty reduction, then examines how estimation rate can trade off with communication rate. Radar capacity is resolution-limited and does not directly map to fundamental communication metrics, while multiple schemes achieve communication–estimation rate inner bounds.

Radar capacity: The resolution-based capacity is not straightforward to trade against fundamental communication metrics because it specifically measures target identifiability.It is a sensing-oriented capacity notion rather than a direct communication-information measure.
Radar capacity: Radar capacity models each range–Doppler–angle resolution cell as a binary information unit that can contain one distinguishable point target.Multiple targets within one resolution unit are treated as a single target, so the measure counts distinguishable target configurations.
Radar capacity: Radar resolution, and thus its noiseless target-identification capacity, is constrained by available temporal, spectral, and spatial resources.For a pulsed radar, range, velocity, and angular resolutions depend on bandwidth, dwell time, antenna configuration, and related physical limits.
Estimation rate: Estimation rate measures cancellation of uncertainty in target parameters per second and can be traded against communication rate in an ISAC receiver.The framework models the target echo as a virtual user and supports inner bounds through sub-band allocation, successive interference cancellation, water filling, and Fisher Information optimization.

D. Tradeoff in S&C Spatial Degrees-of-Freedom · E. Cross-Layer Tradeoff

ISAC exposes a spatial tradeoff because communication benefits from exploiting all channel degrees of freedom, whereas sensing may require suppressing or allocating resources among propagation paths and clutter. At higher layers, sensing accuracy and communication rate compete for wireless resources, with DNN-based accuracy–resource relationships potentially nonlinear and difficult to characterize.

D. Tradeoff in S&C Spatial Degrees-of-Freedom: Communication seeks to exploit all available channel degrees of freedom, while sensing may benefit from treating non-line-of-sight paths as clutter sources.This difference in spatial-resource treatment creates a fundamental S&C tradeoff.
D. Tradeoff in S&C Spatial Degrees-of-Freedom: In the vehicle example, all propagation paths contribute to communication receive power, whereas target sensing is evaluated through the signal-to-clutter-plus-noise ratio.The vehicle acts simultaneously as a communication user and a sensing target in the mmWave ISAC scenario.
D. Tradeoff in S&C Spatial Degrees-of-Freedom: ISAC waveform design must allocate power and other resources across propagation paths to balance communication and sensing performance.Convex optimization techniques may be employed, and the tradeoff extends to scenarios with multiple targets and communication users.
E. Cross-Layer Tradeoff: Cross-layer ISAC performs communication and higher-layer sensing tasks, such as human detection, using commercial wireless devices and sensory-data-trained deep neural networks.The resulting tradeoff is not restricted to physical-layer performance.
E. Cross-Layer Tradeoff: DNN-based sensing makes S&C resource allocation challenging because the relationship between accuracy rate and allocated wireless resources may be mathematically intractable.Accuracy rate denotes the probability that the mobile sensor correctly detects human activities or events.
E. Cross-Layer Tradeoff: A time-division ISAC model interleaves sensing and communication cycles, assigning sensing durations to N targets and round-robin communication durations to K users.The model assumes a total time budget T and constant-power transmission.
E. Cross-Layer Tradeoff: The sensing accuracy is modeled as A = Θ(C), an unknown nonlinear function of the number of sensing cycles, while communication performance is represented by the minimum achievable rate among K users.A classical approximation, Θ(C) ≈ 1 − αC^−β, is proposed to capture the accuracy-function shape.

IV. WAVEFORM DESIGN FOR ISAC … 1) Sensing-Centric Design:

ISAC waveform design ranges from loosely coupled, non-overlapped resource allocation to fully unified waveforms that share temporal, spectral, and spatial resources. Sensing-centric unified designs embed communication information into sensing signals while prioritizing sensing performance, but often trade data rate against sensing preservation.

IV. WAVEFORM DESIGN FOR ISAC: ISAC waveform design spans time-, frequency-, and spatial-division schemes before fully unified designs that overlap temporal, spectral, and spatial resources.The progression reflects increasing integration between sensing and communications, with fully unified designs intended to maximize integration gain.
A. Non-Overlapped Resource Allocation: Time-division separates transmission into radar and radio cycles, frequency-division assigns S&C to different OFDM subcarriers, and spatial-division uses orthogonal antenna groups or beams.Time-division is easiest to implement; frequency allocation depends on channel conditions, S&C KPIs, and transmitter power budget, while spatial separation benefits from MIMO and massive MIMO.
B. Fully Unified Waveform: Fully unified ISAC waveforms follow sensing-centric, communication-centric, or joint design philosophies.These philosophies determine whether sensing, communications, or both functionalities primarily guide waveform construction.
1) Sensing-Centric Design:: Sensing-centric design embeds communication data into sensing waveforms without unduly degrading sensing performance.Embedding can occur in different signal domains, including time-frequency and spatial domains.
1) Sensing-Centric Design:: Chirp-based SCD represents symbols by varying amplitude, start frequency, initial phase, or chirp slope, enabling ASK, FSK, and PSK modulation.A chirp’s bandwidth is B = kT0 and its time-bandwidth product is BT0 = kT0^2.
1) Sensing-Centric Design:: Sidelobe-control SCD encodes data through MIMO radar sidelobe levels while retaining the main beam for target sensing, but rapid channel changes can increase false alarms.In the numerical example, ε = −20dB, while the communication sidelobe alternates between δ0 = −40 dB and δ1 = −20 dB.
1) Sensing-Centric Design:: Index modulation preserves sensing by encoding communication symbols through waveform permutations or agile carrier-frequency and antenna allocations.The basic index-modulation scheme reaches a maximum bit rate of fPRF · log2Nt!, while MAJoRCom reaches fPRF · (Kflog2Mf + Ntlog2Kf); FRaC adds phase modulation for a higher increased bit rate.
1) Sensing-Centric Design:: Because most SCD uses slow-time inter-pulse coding, its bit rate is tied to radar PRF, limiting applications to low or moderate data-rate scenarios.SCD provides favorable sensing performance, but generally does not target high-rate communication requirements.

2) Communication-Centric Design:

Communication-centric design adds sensing to an existing communication waveform while prioritizing communication performance. OFDM-based ISAC can estimate delay and Doppler through decoupled processing, but communication data and unmet radar waveform requirements restrict sensing performance.

Communication-Centric Design: Communication-centric design implements sensing over an existing communication waveform, with communication as the primary guaranteed functionality.Any communication waveform can support monostatic sensing because it is known to the transmitter, but data randomness can degrade sensing.
OFDM-Based ISAC: OFDM is a representative communication-centric waveform because it is compatible with 4G and 5G standards.The communication signal uses data symbols across OFDM symbols and subcarriers.
OFDM-Based ISAC: Element-wise division mitigates random communication data, after which column FFT and row IFFT estimate Doppler and delay.The received samples are organized into a matrix following sampling and block-wise FFT processing.
OFDM-Based ISAC: OFDM processing forms a two-dimensional delay-Doppler profile with a detectable target peak, while decoupling delay and Doppler processing.This decoupling is favorable for radar applications compared with conventional chirp signals.
Limitations: Despite guaranteeing communication performance, OFDM-based ISAC restricts sensing because constant envelope, clutter interference, and correlation properties are insufficiently addressed.Constant envelope supports maximum-power transmission without distortion and maximizes received-echo SNR; clutter resistance and good correlation are also required for radar sensing.

3) Joint Design:

Joint design formulates ISAC waveform synthesis as a scalable tradeoff between communication and sensing, unlike SCD and CCD’s restricted extreme cases. By varying a weighting factor, it smoothly connects sensing-optimal and communication-optimal operation while supporting fast-time data modulation and broader channel applicability.

Joint Design: Joint design addresses the scalability limitations of SCD and CCD by jointly optimizing communication and sensing within one waveform.SCD and CCD implement the two functionalities in restricted extreme cases, whereas joint design is proposed as a more flexible methodology.
Joint Design: The weighting factor ρ ∈[0, 1] controls the relative priority assigned to communication and sensing performance.The communication and sensing priorities are represented by ρ and 1 −ρ, respectively, alongside waveform-shaping constraints such as power and constant-modulus limits.
Joint Design: Varying ρ from 0 to 1 produces a favorable tradeoff whose sensing-optimal and communication-optimal endpoints correspond to CCD and SCD performance.The formulation yields zero-forcing communication at full communication priority and the pure sensing waveform X = X0 at full sensing priority, subject to the same power budget.
Joint Design: With Nt = 16 and K = 4, 6, 8, reducing the number of communication users raises detection probability, while K = 4 increases achievable rate without greatly sacrificing sensing.The numerical study uses CFAR detection for a target at 36°, with false-alarm probability 10−7 and receive SNR −6 dB.
Joint Design: Joint design often outperforms conventional schemes in multiple respects, but it entails higher computational complexity.This summarizes the stated benefit–complexity tradeoff of the JD scheme.
Joint Design: Joint-design waveforms encode communication symbols within pulses, improving data rate and avoiding dependence on a specific communication channel.Each fast-time snapshot represents a communication symbol, unlike SCD’s inter-pulse modulation; the approach applies beyond LoS channels.

V. RECEIVE SIGNAL PROCESSING · A. Fundamental Insights from a Toy Model

ISAC receive processing must jointly decode communications and detect or estimate targets from mixed signals, using cooperation between communication and sensing processors to manage mutual interference. A toy AWGN model shows that optimal processing and constellation design should adapt to interference strength and exploit signal structure.

V. RECEIVE SIGNAL PROCESSING: When communication and sensing signals do not overlap, conventional interference-free processing applies; partial or full overlap creates mutual interference requiring specialized receive processing.The receiver must decode useful communication information while detecting or estimating targets from echoes.
V. RECEIVE SIGNAL PROCESSING: ISAC receivers process mixed communication and echo signals through shared RF hardware before jointly supporting information decoding and target detection, estimation, or recognition.Cooperation between communication and sensing processors facilitates mutual interference cancellation.
A. Fundamental Insights from a Toy Model: The toy AWGN model studies communication in the presence of strong radar interference modeled as a high-amplitude, short-duration pulse alongside a lower-power, narrowband communication signal.From the communication receiver’s perspective, the radar interference has accurately estimable amplitude but randomly fluctuating, difficult-to-track phase.
A. Fundamental Insights from a Toy Model: The model raises two fundamental problems: determining the optimal communication decision region and designing constellations that remain effective under radar interference.The received communication and radar powers are assumed known, while the radar phase is unknown and uniformly distributed over [0, 2π].
A. Fundamental Insights from a Toy Model: At low INR, the ML decoder reduces approximately to Treat-Interference-as-Noise, whereas at high INR it becomes an Interference-Cancellation receiver that pre-cancels radar interference.SER is analyzed across low-, mid-, and high-INR regimes for PAM, QAM, and PSK constellations.
A. Fundamental Insights from a Toy Model: Optimization formulations separately maximize constellation cardinality under SER and normalized-power constraints or minimize SER under the corresponding design constraints.Both problems are generally non-convex and are solved sub-optimally using MATLAB Global Optimization Toolbox with Global Search.
A. Fundamental Insights from a Toy Model: The optimal constellation has a concentric hexagon shape in the low-INR regime and an unequally spaced PAM shape in the high-INR regime for both design criteria.These numerical results show that constellation geometry should adapt to the interference regime.
A. Fundamental Insights from a Toy Model: At mid or high radar INR, receivers can estimate and recover radar interference first, then pre-cancel it while exploiting structural information such as radar’s constant modulus.The paper identifies structural exploitation as a central principle for ISAC receiver design and foreshadows hidden sparsity in S&C signals.

B. ISAC Receiver Design based on Sparsity · VI. COMMUNICATION-ASSISTED SENSING: PERCEPTIVE NETWORK · A. General Framework

The paper presents a sparsity-based iterative receiver that jointly demodulates communication data and recovers interfering radar signals, then frames communication-assisted sensing through frame-level and network-level ISAC in perceptive mobile networks. It classifies sensing by signal direction and transceiver placement, including mono-static, bi-static, and distributed/networked modes.

B. ISAC Receiver Design based on Sparsity: The receiver jointly demodulates a single user’s communication signal and recovers interference from J radar/sensing systems.Its received coded signal comprises communication, coded radar, and noise components.
B. ISAC Receiver Design based on Sparsity: Radar codes are modeled in a low-dimensional subspace using a known waveform dictionary, while communication symbols follow x = H_Cb.Dictionary selection enables radar waveform diversity for interference, clutter, and spectrum-sharing conditions.
B. ISAC Receiver Design based on Sparsity: The iterative receiver reconstructs sparse radar coefficients and sparse communication demodulation errors through an on-grid compressed-sensing problem with l1 penalties.The algorithm refines symbol decisions until successive estimates agree or the maximum iteration count is reached; off-grid variants can use atomic norms.
VI. COMMUNICATION-ASSISTED SENSING: PERCEPTIVE NETWORK: Perceptive networks make sensing a native wireless capability supporting services such as localization, recognition, and imaging through communication-assisted sensing.The paper distinguishes frame-level ISAC, based on communication frames and protocols, from network-level ISAC, based on distributed wireless architectures.
A. General Framework: In a perceptive mobile network, targets may be communication or noncommunication objects, and sensing can use downlink or uplink signals transmitted by a BS or UE.High-mobility examples include tracking a communicating mobile terminal in V2X or UAV networks.
A. General Framework: Sensing operations are categorized as downlink or uplink mono-static, downlink or uplink bi-static, and distributed/networked sensing according to transmitter and receiver locations.Distributed/networked sensing uses multiple transmitters and receivers and requires cooperation between transceivers.
A. General Framework: Uplink mono-static sensing is rarely considered because small UEs usually have limited sensing-receiver capability, although high-computation vehicles may support it.This limitation arises when the UE itself must receive signals reflected from targets.

B. Using 5G-and-Beyond Waveform for Sensing … 1) Feasibility:

5G-and-beyond waveforms offer readily available reference signals for sensing, but practical ISAC faces bandwidth, self-interference, and unknown-payload challenges. Networked sensing architectures, especially C-RAN, provide centralized coordination for flexible sensing modes while trading performance against overhead and hardware costs.

1) Feasibility:: NR synchronization signals, PBCH, DMRS, and Release 16 PRS can support sensing or localization, with PBCH and DMRS additionally open to joint S&C precoding and scheduling optimization.PSS and SSS have fixed structures, whereas PBCH and its DMRS are more flexible.
2) Challenges and Opportunities:: NR bandwidth limits range resolution to 1.5 m in FR1 and 0.375 m in FR2, which may support basic sensing but not high-precision localization.FR1 spans 450 MHz to 6 GHz with up to 100 MHz bandwidth, while FR2 spans 24.25 GHz to 52.6 GHz with up to 400 MHz bandwidth.
2) Challenges and Opportunities:: Mono-static sensing suffers self-interference because the transmitter and receiver share the sensing platform, producing a 100-1000 m black zone for static targets even with one OFDM symbol.Larger black zones arise when both SSB and data payload are used, and practical ADCs cannot handle the resulting signal dynamic range.
2) Challenges and Opportunities:: RF and digital cancellation can mitigate mono-static self-interference, with reported suppression of 100 dB; alternatively, transmit-receive antenna isolation is required.Full-duplex and self-interference cancellation techniques are identified as necessary approaches.
2) Challenges and Opportunities:: Bi-static sensing avoids self-interference, but unknown data payloads cannot be directly used when transmitter-receiver direct links are absent; PBCH-based channel estimation can address this problem.The receiver can estimate the channel matrix from PBCH before handling the payload.
2) Challenges and Opportunities:: NR pilots support communication-channel estimation, after which sensing can estimate target amplitude, delay, Doppler, angle of arrival, and angle of departure from the channel or reconstructed data symbols.Refining the communication channel estimate can provide coordination gain.
C. Using 5G-and-Beyond Network Architecture for Sensing: 5G-and-beyond C-RAN enables centralized networked sensing through coordinated BBUs, distributed RRHs, and fronthaul, supporting information-level and signal-level fusion.Signal-level fusion preserves more sensing information but incurs higher computational, signaling, and hardware costs than information-level fusion.

2) Challenges and Opportunities: … B. Sensing-Assisted Beam Tracking and Prediction

The paper presents sensing-assisted communication as a path to coordinated perceptive networks, focusing on interference exploitation, sensing-enabled beam alignment, and ISAC-based beam tracking and prediction. These methods reduce communication overhead while using sensing information to maintain reliable links in dynamic V2X settings.

2) Challenges and Opportunities:: Inter-RRH interference, typically harmful in communications, can contain useful target information for perceptive networks and support bi-static sensing.Recovering bi-static channel information from other RRHs can help compensate target radar-cross-section fluctuations.
2) Challenges and Opportunities:: Downlink bi-static sensing between RRHs is highlighted as the most promising sensing mode because fronthaul coordination exposes transmitted SSBs and payloads.RRHs connected to the BBU pool can straightforwardly share signals for sensing.
VII. SENSING-ASSISTED COMMUNICATION: Sensing-assisted communication includes CSI estimation from pilots and cognitive-radio spectrum sensing, where detected spectrum availability enables secondary transmission.These examples show communication systems can be assisted by sensing even outside integrated sensing-and-communications designs.
A. Sensing-Assisted Beam Training: Sensor information from GNSS, radar, lidar, or cameras narrows beam-search spaces and reduces training overhead for latency-critical links such as V2X.In V2I, sensing can support beam alignment by restricting the candidate beam pairs before communication training.
A. Sensing-Assisted Beam Training: Radar echoes can estimate the communication-channel covariance and enable RSU precoding without vehicle feedback by exploiting channel reciprocity.The RSU uses the estimated covariance to transmit pilots that facilitate receive beamforming.
B. Sensing-Assisted Beam Tracking and Prediction: Radar-aided tracking reduces training overhead but requires extra radar hardware, while high-mobility V2X additionally motivates beam prediction for fast-changing channels.Prediction can use model-based vehicle kinetics or model-free machine learning for complex traffic and channel conditions.
B. Sensing-Assisted Beam Tracking and Prediction: ISAC iteratively predicts and tracks vehicle state using echoes, preserving V2I quality while eliminating downlink pilots, uplink feedback, and angle quantization errors.Using the entire ISAC block also provides matched-filtering gain equal to the ISAC block length, improving estimation accuracy.
B. Sensing-Assisted Beam Tracking and Prediction: The ISAC beam-tracking framework gains estimation accuracy by using continuous echo-based angle estimation and whole-block matched filtering instead of limited feedback pilots.These gains complement the reductions in downlink and uplink overheads described for communication-only tracking.

C. Sensing-Assisted Resource Allocation

Sensing can support efficient wireless-resource allocation by using vehicles’ kinematic states, environments, and geometrical relationships to jointly address communication and sensing requirements. The section considers sensing-assisted allocation of bandwidth, beamwidth, and power, as well as cell handover, across V2X and broader communication applications.

Sensing-Assisted Cell Handover: Sensing can assist cell handover and other communication applications requiring low overhead, low latency, and efficient resource allocation.The V2I handover example uses dual connections and compares serving- and idle-link receive SINR for continuity of service.
Resource-allocation framework: Sensing-assisted resource allocation uses vehicles’ kinematic states, driving environments, and geometrical relationships while accounting for both S&C performance.The framework is especially relevant to simultaneous service of multiple vehicles in V2X scenarios.
Bandwidth Allocation: Bandwidth allocation must balance communication spectral efficiency or user QoS, mutual-interference mitigation, and sensing range-resolution requirements.Different vehicles may impose distinct communication-rate and sensing-resolution constraints under a shared spectrum budget.
Beamwidth Allocation: Beamwidth allocation jointly affects communications and sensing: narrow beams suit faraway point-like vehicles, whereas nearby extended vehicles require wider beams.Narrow beams provide communication beamforming gain and angular sensing resolution for distant point-like targets.
Power Allocation: Power allocation must jointly address communication rate and estimation accuracy because classical water-filling maximizes communication rate but does not minimize estimation errors.V2X services may simultaneously require high-throughput communication and high-accuracy localization, with CRB and communication rate among the relevant metrics.

D. Sensing-Assisted PHY Security … 3) ISAC With UAVs:

The paper frames ISAC as a foundation for 6G networks that integrates sensing, communications, security, computation, reconfigurable surfaces, and aerial platforms. It highlights mutual assistance between sensing and communications while identifying new cross-domain tradeoffs and future technology integrations.

D. Sensing-Assisted PHY Security: ISAC creates new PHY-security challenges because probing signals carry data that malicious radar targets may eavesdrop on, while communication-link detection can expose network identity and location.Communications-only countermeasures that suppress power toward eavesdroppers can severely degrade ISAC sensing performance.
D. Sensing-Assisted PHY Security: Sensing can make PHY-security solutions practical by detecting eavesdroppers’ targets and channels or directions, while introducing tradeoffs among communications, sensing, and security.Security is presented as an additional requirement for future wireless networks.
1) ISAC Meets Edge Intelligence:: Integrating ISAC with edge intelligence intensifies spectrum bottlenecks because limited resources must support both AI-model exchange and radio sensing.The integration also creates sensing–communication–computation tradeoffs, requiring joint design around ultimate AI-task goals.
2) ISAC Supported By Reconfigurable Intelligent Surface (RIS):: RIS can improve wireless efficiency, while ISAC can estimate user-related parameters to overcome costly RIS-related CSI acquisition and facilitate passive beamforming.RIS elements independently control reflecting amplitudes and phases, but lack signal-processing capability for direct CSI acquisition.
3) ISAC With UAVs:: UAVs can serve as ISAC sensing targets, communication users, relays, aerial base stations, or access points within integrated terrestrial-and-aerial networks.This creates a research area centered on the interplay between UAVs and ISAC.
3) ISAC With UAVs:: ISAC-enabled cellular networks can detect suspicious UAVs for cyber-physical protection and localize connected UAVs during communication to support beam tracking and wireless resource management.The sensing measurements also provide related channel parameters for network operation.
B. Interplay Between ISAC and Other Emerging Technologies:: For emerging technologies, the paper identifies potential ISAC combinations with LEO satellite networks, THz communications and sensing, digital twins, and OTFS modulation.These directions are noted but not elaborated because of the page limit.
A. Summary: The paper envisions ISAC as a foundation for 6G’s new air interface and a bond connecting the physical and cyber worlds through pervasive sensing, connectivity, and intelligence.The broader vision is supported by the paper’s review of ISAC history, gains, applications, industrial progress, and standardization.