PaperlayerView exact revision on arXiv

Source-linked AI summary

Rotatable Antenna-Empowered Wireless Networks: A Tutorial

Beixiong Zheng, Qingjie Wu, Xue Xiong, Yanhua Tan, Tiantian Ma, Qi Dai, Weihua Zhu, Changsheng You, Xiaodan Shao, Lipeng Zhu, Jie Tang, Robert Schober, Kai-Kit Wong, Rui Zhang

arXiv:2603.25559v3cs.ITcs.ETeess.SP

TL;DR

Fixed-antenna scaling is constrained by hardware complexity, energy consumption, static orientations, and dynamic propagation conditions, motivating more adaptable spatial architectures. This tutorial develops a unified account of RA architectures, models, optimization, channel acquisition, deployments, and prototypes, with experiments reporting approximately 7 dB SNR improvement over fixed antennas. It also identifies practical overheads, estimation limits, and deployment-specific challenges that constrain RA systems.

  • Problem

    Fixed-antenna scaling faces hardware complexity, energy consumption, diminishing returns, and mismatch with dynamic propagation environments, while RA requires practical treatment beyond ideal orientation control.

  • Method

    The tutorial unifies RA rotation and channel models, communication and sensing optimization, multi-view channel acquisition, practical architectures, deployment strategies, prototypes, and experimental evidence.

  • Results

    Approximately 7 dB SNR improvement is observed for sensing-assisted RA implementations compared with a conventional fixed-antenna system.

  • Takeaways & Limitations

    RA provides orientation-based spatial adaptation for communication and sensing without extra antenna resources or deployment space.

Abstract

from arXiv · show

Non-fixed flexible antenna architectures, such as fluid antenna system (FAS), movable antenna (MA), and pinching antenna, have garnered significant interest in recent years. Among them, rotatable antenna (RA) has emerged as a promising technology for enhancing wireless communication and sensing performance through flexible antenna orientation/boresight rotation. By enabling mechanical or electronic boresight adjustment without altering physical antenna positions, RA introduces additional spatial degrees of freedom (DoFs) beyond conventional beamforming. In this paper, we provide a comprehensive tutorial on the fundamentals, architectures, and applications of RA-empowered wireless networks. Specifically, we begin by reviewing the historical evolution of RA-related technologies and clarifying the distinctive role of RA among flexible antenna architectures. Then, we establish a unified mathematical framework for RA-enabled systems, including general antenna/array rotation models, as well as channel models that cover near- and far-field propagation characteristics, wideband frequency selectivity, and polarization effects. Building upon this foundation, we investigate antenna/array rotation optimization in representative communication and sensing scenarios. Furthermore, we examine RA channel estimation/acquisition strategies encompassing orientation scheduling mechanisms and signal processing methods that exploit multi-view channel observations. Beyond theoretical modeling and algorithmic design, we discuss practical RA configurations and deployment strategies. We also present recent RA prototypes and experimental results that validate the practical performance gains enabled by antenna rotation. Finally, we highlight promising extensions of RA to emerging wireless paradigms and outline open challenges to inspire future research.

I. INTRODUCTION

Wireless networks face increasingly scarce spectrum and diminishing returns from scaling fixed antennas, whose static orientations cannot adapt to dynamic propagation environments. The tutorial presents rotatable antennas as a flexible approach that adds orientation-based spatial DoFs for communication, sensing, and diverse deployments.

  • A. Background and Motivation: Spectrum scarcity and the diminishing returns of antenna scaling limit conventional approaches to increasing wireless-network capacity.Massive and extremely large-scale MIMO also incur high RF-chain cost, circuit power, signal-processing complexity, and hardware expense.
  • A. Background and Motivation: Fixed-antenna systems keep antenna positions and orientations static after deployment, creating mismatch with dynamic wireless propagation and limiting spatial adaptability.This constraint prevents fixed architectures from fully exploiting available spatial DoFs.
  • B. What Is RA and How It Works: RA adds spatial DoFs by independently rotating antenna orientations or boresights in three-dimensional space without changing antenna positions.RA can concentrate radiated energy toward desired directions and complement conventional beamforming over fixed antennas.
  • B. What Is RA and How It Works: RA can be implemented mechanically or electronically, with co-designed architectures combining wide angular adjustment and rapid boresight control.Mechanical approaches offer wide angular ranges with milliwatt-level power and millisecond response, while electronic approaches provide faster adjustment and compatibility with existing systems.
  • B. What Is RA and How It Works: Antenna rotation supports adaptive interference circumvention, beam focusing, coverage extension, multi-target sensing, high-resolution sensing, and richer target-feature extraction.Sensing benefits arise from boresight tracking and spatial scanning that observe targets from diverse directions, even with few antennas.
  • C. Historical Development: RA developed from mechanically rotating antennas and electronically steered radiation toward a rotation-centric architecture distinct from position-reconfigurable MA and FAS.The tutorial positions RA as a balance between flexibility and complexity for space-, payload-, cost-, power-, and control-constrained deployments.
  • D. Potential of RA Deployment in Future Wireless Networks: RA is applicable across space, air, and ground segments, including aerial or satellite platforms, dense indoor environments, smart factories, and immersive applications.Its directional adaptability supports spatially varying service demands and flexible alignment with users, devices, or machines.

E. Objective, Contribution, and Organization

The tutorial establishes a unified theoretical and practical foundation for RA-enabled wireless networks, addressing a research gap in existing flexible-antenna overviews. It covers RA fundamentals, optimization, channel acquisition, configurations, prototypes, and future directions.

  • Objective: The tutorial addresses the limited transferability of existing flexible-antenna studies to RA systems by developing a dedicated overview and framework.Existing works commonly focus on specific reconfiguration dimensions and modeling assumptions.
  • Contributions: Its framework includes general antenna/array rotation models and channel models capturing boresight-dependent gain and polarization effects.The paper also develops near-field and extended propagation models for RA systems.
  • Contributions: The tutorial surveys rotation optimization across communication and ISAC settings, including single-user, multi-user, MIMO, and wideband systems.The covered designs use different performance metrics across representative RA-aided scenarios.
  • Contributions: It reviews RA channel estimation and acquisition through fixed or dynamic orientation estimation and multi-view signal processing.Orientation scheduling is included among the discussed estimation strategies.
  • Contributions: It compares practical RA configurations, prototypes, implementation strategies, experimental results, and open research challenges.Configuration comparisons include sparse versus non-sparse arrays, continuous versus discrete rotation, and distributed versus centralized deployment.
  • RA Fundamentals: The rotation model represents antenna orientation using three angles and two non-parallel vectors that support gain-only and polarization-aware channel models.The pointing vector determines directional gain, while the reference vector characterizes polarization direction.

2) Array Rotation:

Array rotation changes both antenna orientations and their positions relative to the array center. Joint array and individual-antenna rotation therefore requires modeling both rotation levels and their combined energy consumption.

  • Array Rotation: Rotating the entire array changes antenna orientations and spatial positions relative to the array center.The post-rotation position is defined from the array rotation center and each antenna’s initial position.
  • Array Rotation: Without independent antenna rotation, identically oriented antennas retain common pointing and reference vectors after array rotation.This case corresponds to a shared array rotation with no additional per-antenna orientation change.
  • Array Rotation: With simultaneous independent antenna rotation, both array-level and individual-antenna rotations determine each antenna’s pointing and reference vectors.The resulting orientation model accounts for the combined transformations.
  • Energy Consumption: Joint rotation energy equals the sum of array-level actuation energy and individual antenna-level rotation energies.The same mechanical-energy model applies after substituting equivalent angular displacement and the array’s hardware coefficient.

B. Channel Model

The RA channel framework begins with directional antenna gain models and incorporates orientation-dependent gain into near-field LoS channels, with extensions to far-field and broader propagation scenarios. Representative cosine and 3GPP patterns capture different antenna directivity characteristics.

  • Channel Framework: The channel framework models RA systems using directional gain patterns, near-field LoS propagation, far-field approximation, multipath, and wideband extensions.The framework begins with radiation-energy models before constructing the channel expressions.
  • Directional Gain Models: The cosine pattern models a narrow mainlobe with negligible sidelobes using a directivity parameter ρ.Its maximum gain is Gcos_max = 2(2ρ+1), and larger ρ produces higher boresight gain with a narrower mainlobe.
  • Directional Gain Models: The 3GPP element model represents practical directional gain using maximum gain, front-to-back attenuation, beamwidths, and vertical sidelobe suppression.Its horizontal and vertical components are expressed in dB and generally produce an irregular pattern.
  • Directional Gain Models: The framework can substitute analytic, measured, or full-wave simulated radiation patterns according to antenna structure, materials, feeding, and frequency.Examples include dipoles, monopoles, horn antennas, and microstrip patches.

2) Near-Field LoS Channel Model:

The near-field LoS model computes orientation-dependent gain from each RA to a user using local incident/departure angles and propagation geometry. Far-field models simplify this when link distance is sufficiently larger than array aperture.

  • Near-Field Geometry: The effective gain from each RA depends on the user direction relative to the antenna’s current pointing and reference orientation.The user-to-RA direction is obtained from their positions and used to derive local incident/departure angles.
  • Near-Field Gain: Under the rotationally symmetric cosine model, gain increases as the boresight approaches the user direction and is maximized when f⊥,n = qU,n,k.This gain-only expression describes directional variation without polarization mismatch.
  • Polarization: When polarization mismatch is non-negligible, the reference vector must be retained to evaluate the polarization matching factor.The directional-gain-only model is appropriate when polarizations are aligned or compensated.
  • Near-Field Channel: The near-field LoS channel combines orientation-dependent gain with distance-dependent propagation coefficients for each RA-user link.The channel is stacked across all RAs, whose orientations are represented by an orientation matrix.
  • Far-Field Approximation: The far-field approximation treats directions and propagation coefficients as approximately identical across RAs when link distance greatly exceeds array aperture.The resulting model uses a common user direction and propagation coefficient across the array.

4) Multipath Channel Model:

The multipath channel model represents line-of-sight and clustered non-line-of-sight propagation, then extends it to wideband and polarization-aware settings. Antenna rotation affects channel amplitude through both directional gain and polarization matching.

  • Multipath propagation: Clustered scattering models the non-line-of-sight channel using scatterer positions, directional gains, radar cross sections, and path distances.The overall channel is obtained by superimposing the LoS and NLoS components.
  • Wideband channel: OFDM modeling represents wideband propagation through frequency-selective space-time impulse responses and space-frequency responses across subcarriers.The formulation uses bandwidth B, L subcarriers, propagation delays, and subcarrier spacing.
  • Polarization effects: Polarization-aware modeling links each RA’s reference vector to its in-plane polarization axis and projects that direction onto the plane orthogonal to propagation.The effective electric-field direction depends jointly on antenna polarization and user direction.
  • Polarization effects: Rotation-induced polarization mismatch reduces received power when the incoming electric-field direction does not align with the receive polarization direction.The mismatch is characterized through a polarization matching gain incorporated into the LoS channel coefficient.
  • Polarization effects: Antenna orientation must balance directional gain and polarization matching because rotation changes both contributions to channel amplitude.The model can also extend to cross-polarized antenna pairs through element-wise linear-polarization modeling.

7) Extensions to Other Channel Models:

The paper extends the geometry-based RA framework to generalized system models, stochastic propagation, practical hardware responses, and resource-constrained optimization. These extensions preserve orientation-dependent modeling while accounting for uncertainty and implementation constraints.

  • Generalized channel models: The baseline BS-side RA model generalizes to MIMO links with RA arrays at users and supports direct orientation optimization.The same framework can represent antenna positions, orientations, and propagation-dependent responses in broader link configurations.
  • Stochastic models: Stochastic extensions treat user locations, scatterer positions, and small-scale fading coefficients as random variables, including Rician LoS and spatially correlated NLoS components.The associated statistics are often modeled as stationary or quasi-static, although mobility can challenge that assumption.
  • Practical responses: Ideal channel models may deviate from practical compact-array behavior because rotation changes mutual coupling, radiation patterns, phase centers, impedance, and polarization.These effects make the actual response orientation-dependent.
  • Practical responses: In-situ calibration measures or simulates radiation patterns for each RA orientation after integration with the final mechanical platform.The calibrated responses can replace ideal directional-gain patterns in the channel model.
  • Optimization framework: The generic optimization framework jointly selects rotation variables and system resources under rotation, hardware, and operational constraints.Its utility can represent communication metrics such as rate or outage and sensing metrics such as CRB, detection probability, or estimation accuracy.

3) Other Constraints:

RA optimization is constrained by hardware, resource, and implementation limits, but orientation control can provide substantial array-gain benefits. In the considered MISO setting, gains scale linearly only before the allowable rotation range causes saturation.

  • Practical constraints: RA optimization must account for implementation cost, energy, latency, computational complexity, hardware imperfections, imperfect CSI, finite rotation speed, and angular resolution.Ideal-response optimization can become suboptimal when practical impairments cause model mismatch, degraded coherent combining, or interference leakage.
  • Practical constraints: Two-timescale control stores offline orientations or codebooks for representative states and selects or interpolates them online from observations.This design targets more feasible real-time RA control.
  • MISO optimization: Each RA orientation affects SNR through the projection of its pointing vector onto the user direction, so optimal boresights align as closely as rotation limits permit.Under the cosine pattern model, the orientation problem becomes a per-antenna projection maximization.
  • MISO scaling: The closed-form MISO result has two regimes: all RAs can align within the rotation range up to a threshold, while additional antennas eventually exceed that range.The maximum SNR depends on user span angle and allowable rotation range.
  • MISO scaling: In the fully alignable regime, received SNR increases approximately linearly with N; beyond the threshold, its growth gradually diminishes and approaches saturation.A larger allowable rotation range produces a higher asymptotic SNR.
  • MISO evaluation: In the 2.4-GHz, d = 15 m simulation, RA-enabled MISO achieves up to 5 dB higher received power than fixed antennas when N ≤100.The advantage arises because each RA independently steers its boresight toward the user.
  • Distance effects: Near-field angular variation lets RA orientation exploit antenna-dependent user directions, while optimized orientations converge as distance increases and angular spread diminishes.The RA system nevertheless maintains higher received power in the considered distance comparison.

B. RA-Enabled MIMO System

RA-enabled MIMO jointly optimizes transmit and receive orientations with covariance allocation, while multi-user systems distribute boresights across user-specific propagation directions. These added spatial DoFs improve channel conditioning, multiplexing, and max-min performance under directional propagation.

  • MIMO benefits: Joint transmit- and receive-side orientation adjustment can align dominant paths, reduce inter-path correlation, and produce a higher-rank, better-conditioned MIMO channel.The resulting channel improvement strengthens spatial multiplexing and increases achievable channel capacity.
  • MIMO optimization: RA-enabled MIMO capacity optimization jointly selects transmit orientations, receive orientations, and the transmit covariance matrix under power and rotation constraints.The resulting problem is highly non-convex because orientations reshape the channel eigenstructure while covariance controls eigenmode allocation.
  • MIMO optimization: Alternating optimization addresses the MIMO coupling by updating covariance through eigenmode transmission and water-filling, then alternating transmit- and receive-side orientation updates.This separates the covariance and orientation updates while retaining their joint objective.
  • Multi-user operation: In multi-user systems, different RAs can steer toward different user directions, while multipath-aware designs allocate orientations across dominant paths.This contrasts with pointing all antennas toward one direction or indiscriminately toward all users and scatterers.
  • Multi-user optimization: The multi-user max-min SINR problem jointly optimizes receive beamforming and RA pointing vectors, typically through alternating updates with ZF or MMSE receivers.The coupled objective is highly non-convex.
  • Multi-user evaluation: RA max-min achievable rate increases with directivity factor ρ, whereas fixed-antenna performance decreases for ρ ≥1 because narrower mainlobes penalize off-axis users.Random orientation is inferior because it does not strategically balance directional gains across multipath channels.

D. RA-Enabled Wideband System

RA-enabled wideband systems jointly optimize antenna pointing and subcarrier assignment to improve frequency-selective channel quality and achievable sum-rate. RA orientation also supports wideband ISAC trade-offs by adapting directional gain to communication and sensing requirements.

  • Wideband RA principles: RAs adjust boresights toward dominant paths, providing spatial selectivity that mitigates frequency-selective fading and improves effective channel responses across subcarriers.The mechanism attenuates weaker or delayed multipath components arriving from other directions.
  • Wideband optimization: The wideband sum-rate problem jointly optimizes RA pointing vectors and binary subcarrier assignments under per-user power and single-user-per-subcarrier constraints.Alternating optimization addresses the resulting non-convex coupling between orientation and allocation.
  • Wideband results: The RA-enabled wideband system consistently achieves the highest sum-rate as the number of subcarriers increases.Larger L reduces relative cyclic-prefix overhead and gives RA systems more subcarrier-allocation flexibility.
  • RA-enabled ISAC: In ISAC, independently oriented RAs achieve higher minimum echo power than baseline schemes and enlarge the communication–sensing trade-off region.The gain comes from reconfiguring the directional pattern in response to the environment and ISAC requirements; the gap over array-wise rotation widens as Rmin increases.
  • Practical limitations: RA optimization methods based on AO/BCD and SCA can be computationally demanding and initialization-sensitive in large-scale or rapidly time-varying systems.Dominant-direction initialization, warm starts, and two-timescale control are suggested to reduce overhead and improve convergence.

IV. RA CHANNEL ESTIMATION/ACQUISITION

RA channel acquisition exploits orientation-dependent observations while preserving the propagation geometry, making orientation scheduling central to estimation design. The main strategies differ in whether antenna orientations remain fixed or vary during training.

  • Channel-acquisition basis: RA rotation changes the radiation pattern but not propagation geometry, including path loss, AoAs/AoDs, or delays.After antenna characterization, the orientation-dependent directional gain can be treated as known.
  • Estimation strategies: Channel acquisition is classified into fixed-orientation and dynamic-orientation strategies according to whether rotation is used during training.The two approaches trade implementation overhead against orientation diversity and estimation accuracy.
  • Fixed-orientation estimation: Fixed-orientation estimation uses a single-view observation and requires an additional configuration period before data transmission.It has low control overhead and remains compatible with conventional fixed-antenna estimation methods.
  • Dynamic-orientation estimation: Dynamic-orientation estimation reconfigures boresights across training blocks to collect multiple orientation-dependent pilot observations.These observations can improve parameter identifiability and estimation accuracy, but require additional pilot and switching overhead.
  • Channel modeling: Near-field RA channel models represent directional gains and distance-dependent array responses for line-of-sight and scattered paths, while far-field responses simplify the distance variation.In the far field, amplitude variation becomes negligible and phase variation is approximately linear across antennas.

1) Dynamic-Orientation Channel Estimation:

Dynamic-orientation channel estimation gathers multi-view pilot observations by changing RA orientations during training and coherently fusing them. Block-wise scanning reduces switching overhead while retaining orientation diversity.

  • Orientation scheduling: The dynamic strategy changes the RA orientation matrix across training slots or blocks to acquire orientation-diverse pilot observations.Block-wise scanning keeps orientations fixed for Tb slots and changes them across M blocks, with Ta = MTb.
  • Implementation trade-offs: Frequent orientation switching introduces control latency, calibration and feedback overhead, and higher multi-view fusion complexity.Efficient scanning strategies and low-complexity estimators are therefore needed for practical dynamic acquisition.
  • Multi-view processing: Multi-view measurements can be coherently fused because key channel parameters remain invariant to antenna orientation.This improves estimation accuracy by exploiting the additional spatial degrees of freedom provided by rotation.
  • Fixed-orientation comparison: Fixed-orientation estimation keeps the orientation matrix unchanged and is equivalent to conventional fixed-antenna estimation with a deterministic radiation pattern.Its low complexity and negligible orientation-control overhead come with limited orientation diversity, sensing coverage, and estimation accuracy in rich multipath.
  • Estimation benefits: Dynamic orientation can improve parameter estimation by increasing directional gain and enriching orientation-dependent observations.Reported applications include sequential scanning for improved AoA sensing and multi-configuration observations that alleviate sparse-array ambiguity.

B. Channel Estimation for Different RA System Setups

RA channel acquisition differs across single-RA and multi-RA setups: single-RA systems obtain spatial diversity sequentially, whereas multi-RA arrays collect spatial observations in parallel. The resulting gains are balanced against view, control, and processing constraints.

  • Single-RA setup: Single-RA transceivers use boresight sweeping to obtain multi-perspective observations with lightweight hardware and a single RF chain.They can operate through passive listening or active sensing, jointly processing measurements accumulated across orientation blocks.
  • Single-RA limitations: Single-RA acquisition has limited view diversity and parameter resolvability because switching latency and channel variation bound the number of orientations within a coherence block.Dense sampling of weak paths increases pilot overhead and training latency in rich-scattering environments.
  • Multi-RA setup: Multi-RA arrays acquire N-dimensional spatial snapshots in parallel and concatenate them across controlled orientation schedules.The resulting spatio-temporal observation matrix jointly captures spatial array responses and temporal evolution induced by rotation.
  • Multi-RA trade-offs: Increasing N provides more information for multipath resolution and parameter estimation through high-dimensional multi-view observations.This advantage requires low-overhead orientation control and advanced spatio-temporal processing to manage configuration growth and inference complexity.

C. Signal Processing Methods for RA Channel Estimation

RA channel estimation uses multiple signal-processing paradigms, including ML, subspace, compressed sensing, beam training, and learning-based methods. These methods exploit sparse propagation structure and orientation-diverse multi-view observations, while balancing estimation accuracy, training requirements, and computational complexity.

  • ML-Based Estimation: ML estimation jointly optimizes angular parameters and propagation coefficients from multi-view observations, with orientation schedules affecting measurement conditioning.Repeated or similar orientations can correlate measurements and require more pilots, whereas diverse angular coverage improves the sensing matrix.
  • ML-Based Estimation: ML-based estimation can achieve high accuracy and asymptotically approach the CRB, but its joint optimization is generally high-dimensional, non-convex, and computationally expensive.Alternating minimization methods such as block coordinate descent are motivated as more efficient alternatives.
  • Subspace-Based Estimation: Subspace methods exploit the low-rank structure induced by finitely many dominant propagation paths, using covariance decomposition and MUSIC or ESPRIT to estimate angular parameters.Block-wise spectra from different RA orientations can be fused, after which propagation coefficients are estimated and the channel is reconstructed parametrically.
  • Compressed-Sensing-Based Estimation: Compressed-sensing methods discretize the angular and distance domain and recover sparse channel parameters from reduced pilot observations using algorithms such as OMP and CoSaMP.The orientation-dependent sensing matrix influences mutual coherence, conditioning, and recovery accuracy, making orientation-aware design important.

4) Beam Training-Based Estimation:

Beam training estimates implicit CSI by searching candidate beam and orientation configurations, while learning-based methods map multi-view pilot observations to compact channel representations. Both approaches use RA orientation diversity, but practical designs must address codebook overhead, model complexity, and channel evolution.

  • Beam Training-Based Estimation: Beam training selects the beam and orientation codeword pair that maximizes received power or SNR, avoiding explicit recovery of geometric channel parameters.This makes it attractive when low-complexity link alignment is the main objective.
  • Beam Training-Based Estimation: Radiation-pattern-aware codebooks jointly encode beamforming directions and antenna orientations, with near-field and wideband settings potentially adding focal-distance resolution.Hierarchical coarse-to-fine scanning reduces the overhead of exhaustive codeword searches.
  • Learning-Based Estimation: Learning-based estimation maps stacked multi-view pilot observations to compact channel parameters and reconstructs CSI instead of directly estimating the full channel.Different orientation configurations enrich training data, increasing view diversity for feature learning.
  • Learning-Based Estimation: Learning-based methods can provide robustness to measurement or model impairments and low-complexity online inference after offline training.Once trained, acquisition requires only a forward pass rather than iterative model-based optimization.
  • Practical Considerations: Snapshot-based training can become inaccurate when orientation switching approaches the channel coherence time, causing channel aging, estimation mismatch, and degraded acquisition.Previously estimated dominant directions may still preserve some directional gain when geometric parameters vary more slowly than small-scale fading.
  • Practical Considerations: Mechanical rotation and electronic rotation offer complementary trade-offs, with hybrid architectures combining wide-angle coarse steering and low-latency fine adjustment.Mechanical control handles infrequent large-angle updates, while electronic control compensates for residual pointing errors and supports fast tracking.
  • Practical Considerations: Increasing discrete rotation resolution improves steering accuracy but enlarges codebooks and feedback overhead, while fewer levels reduce cost at the expense of performance.Discrete rotation also introduces optimization variables that are harder to handle than continuous ones.

C. Sparse vs. Non-Sparse Array

RA array structure and deployment shape aperture, radiation patterns, spatial diversity, synchronization demands, and practical performance. Sparse arrays combine enlarged-aperture effects with orientation reconfigurability, while prototypes and commercial products demonstrate implementation feasibility and measured gains.

  • Sparse vs. Non-Sparse Array: Array geometry determines aperture and spatial sampling, thereby shaping mainlobe width, sidelobe level, grating-lobe behavior, and achievable sensing and communication performance.RA structure must balance hardware cost, complexity, deployment space, and system requirements.
  • Sparse vs. Non-Sparse Array: Sparse RA arrays enlarge aperture without increasing antenna count, producing sharper beams and finer spatial resolution while using rotation to regulate sidelobes and grating lobes.This joint design supports orientation diversity for sensing and interference shaping for communication.
  • Deployment Strategies: Centralized deployment favors coherent beamforming and tight synchronization, whereas distributed deployment provides location diversity but requires stricter synchronization, CSI exchange, and calibration.Deployment choice also depends on cost, user or target distribution, space constraints, and propagation conditions.
  • Prototype Validation: 7 dB SNR improvement was observed over a fixed-antenna system in 5.8 GHz sensing-assisted RA experiments using 16-QAM, 10 dBm transmit power, and a 2 Mbps data rate.The experiments also produced a clearer 16-QAM constellation, supporting the practicality of sensing-assisted boresight adjustment.
  • Prototype Validation: RA arrays were demonstrated for indoor high-gain directional communication and low-altitude ISAC with real-time sensing and communication involving a UAV.The array uses target perception and boresight adjustment to concentrate signal power and estimate incoming directions.
  • Prototype Validation: 171.25 USD was reported as the total hardware cost of an RA-array prototype including directional antennas, servo gimbals, an FPGA, and control modules.The reported cost supports the stated practicality and scalability of RA architectures for dense low-altitude ISAC deployments.
  • Prototype Validation: Mechanical RA prototypes remain constrained by motor response time, angular resolution, and control accuracy.These proof-of-concept limitations define practical boundaries for the demonstrated performance gains.
  • Related Commercial Products: Commercial products implement related orientation control through motorized antennas or electronic switching among radiating elements.Examples include TP-Link’s Archer AXE200 Omni and Huawei smart-antenna solutions, extending RA-like concepts beyond research prototypes.

VII. EXTENSIONS AND FUTURE DIRECTIONS

RA extensions span low-altitude ISAC, cognitive radio, physical-layer security, cell-free MIMO, SWIPT, IRS, and cross-disciplinary systems. These directions add application opportunities while raising coordination, channel-acquisition, actuation, mobility, and security challenges.

  • A. Low-Altitude ISAC: RA arrays improve low-altitude ISAC by reconfiguring boresights for 3D coverage, target tracking, and sensing under aerial mobility.UAV-mounted RAs can jointly optimize trajectory and antenna orientation to establish line-of-sight links and illuminate blind spots.
  • A. Low-Altitude ISAC: Low-altitude RA deployments still require coordinated ground-air operation, fast actuation, and high-precision alignment for continuous coverage and reliable tracking.High-speed UAV trajectories and rapidly varying channels impose stringent real-time control requirements.
  • B. Cognitive Radio (CR) Systems: RA-aided cognitive radio can support spectrum sensing and opportunistic access, but uncertain primary-user location information and dynamic reoccupation demand rapid orientation reconfiguration.Boresight control must maintain secondary-user connectivity while keeping interference below regulatory thresholds.
  • C. Physical Layer Security (PLS): Joint transmit-beamforming and RA-angle optimization strengthens legitimate links and suppresses leakage, while statistical-CSI orientation optimization supports non-real-time secure deployment.These approaches improve achievable secrecy rate or average secrecy capacity under different channel-information and update constraints.
  • C. Physical Layer Security (PLS): RA-equipped adversaries can improve interception or jamming, motivating joint antenna-orientation, transmission, and interference-management designs.Proposed defenses include artificial-noise steering, secure beamforming, and spatial null formation toward adversarial directions.
  • D. Cell-Free MIMO Networks: In cell-free MIMO, RA supports AP-user pairing, cooperative precoding, and spatially selective cooperation that can reduce unnecessary coordination and signaling overhead.The resulting design space includes single-RA APs and multi-RA array APs, with implications for fronthaul-efficient control.
  • D. Cell-Free MIMO Networks: Cell-free RA systems face tightly coupled orientation, precoding, association, fronthaul, and imperfect-CSI constraints, making scalable joint optimization nontrivial.Future work targets distributed boresight control, joint AP-user association and orientation optimization, and robust limited-signaling designs.
  • E. SWIPT: RA-aided SWIPT must account for orientation-update accuracy, update frequency, actuation energy, and latency when optimizing net energy efficiency.Rotation overhead can shift the rate-energy operating point, particularly for low-power IoT devices.
Loading 2603.25559v3…