PaperlayerView exact revision on arXiv

Source-linked AI summary

An Overview on Integrated Localization and Communication Towards 6G

Zhiqiang Xiao, Yong Zeng

arXiv:2006.01535v1eess.SP

TL;DR

The paper addresses the need to combine localization and communication for 6G because existing localization can be too coarse for emerging applications and separate designs underuse shared infrastructure and resources. It provides a tutorial survey of localization fundamentals, ILAC interactions, co-design techniques, and aerial-ground 3D networks. The paper concludes with an envisioned 6G architecture and future directions including AI-enabled resource management, spectrum and waveform design, and shared receiver processing.

  • Problem

    Existing localization and communication research is largely separate, while 6G applications require accurate location information and coordinated use of network infrastructure and radio resources.

  • Method

    The article gives a tutorial overview covering wireless localization basics, 5G localization recommendations, enabling technologies, location-aided communication, ILAC co-design, and aerial-ground integrated networks.

  • Results

    The article presents architectures and techniques for integrating localization and communication across network layers and for current 2D and future 3D networks.

  • Takeaways & Limitations

    ILAC can use location information to assist communication and communication technologies to improve localization, supporting dual-purpose use of wireless infrastructure and radio resources.

Abstract

from arXiv · show

While the 5G cellular system is being deployed worldwide, researchers have started the investigation of the 6G mobile communication networks. Although the essential requirements and key usage scenarios of 6G are yet to be defined, it is believed that 6G should be able to provide intelligent and ubiquitous wireless connectivity with Tbps data rate and sub-millisecond latency over 3D network coverage. To achieve such goals, acquiring accurate location information of the mobile terminals is becoming extremely useful, not only for location-based services but also for improving wireless communication performance in various ways such as channel estimation, beam alignment, medium access control, routing, and network optimization. On the other hand, the advancement of communication technologies also brings new opportunities to greatly improve the localization performance, as exemplified by the anticipated centimeter-level localization accuracy in 6G by ultra massive MIMO and mmWave technologies. In this regard, a unified study on integrated localization and communication (ILAC) is necessary to unlock the full potential of wireless networks for the best utilization of network infrastructure and radio resources for dual purposes. While there are extensive literatures on wireless localization or communications separately, the research on ILAC is still in its infancy. Therefore, this article aims to give a tutorial overview on ILAC towards 6G wireless networks. After a holistic survey on wireless localization basics, we present the state-of-the-art results on how wireless localization and communication inter-play with each other in various network layers, together with the main architectures and techniques for localization and communication co-design in current 2D and future 3D networks with aerial-ground integration. Finally, we outline some promising future research directions for ILAC.

I. INTRODUCTION

Future 5G/6G applications require accurate, low-latency localization across interactive, urban, indoor, transportation, and factory scenarios. This motivates integrated localization and communication, which the article surveys as a way to use wireless infrastructure and radio resources for both functions.

  • Motivation: Current GNSS, WLAN, and cellular systems may provide only meter-level accuracy in cluttered environments, below the centimeter-level needs of emerging applications.The paper identifies accurate localization as important for intelligent interactive networks, smart cities, and automatic factories.
  • Intelligent Interactive Networks: Wireless XR requires localization accuracy from 1 cm to 10 cm and measurement delay typically below 20 ms, depending on the usage scenario.Tracking accuracy and measurement delay affect XR information transmission and user experience.
  • Smart Indoor Services: Indoor localization is challenged by severe NLoS propagation and location-privacy requirements, which differ across public and personal devices.NLoS propagation can significantly degrade localization accuracy, while the information requiring protection depends on the device context.
  • Smart Transportation: Smart transportation needs accurate, wide-coverage, and robust localization in highly mobile settings for applications including autonomous driving and V2V communications.Autonomous driving uses relative vehicle-obstacle distances for real-time 3D mapping, while V2V localization can improve communication performance.
  • Future Network Design: Future networks are expected to integrate communication, computing, control, localization, and sensing while managing radio resources according to mobile-terminal locations.The article frames this direction as part of 3CLS and self-sustaining networks for maintaining key performance indicators.
  • Article Scope: The article surveys localization basics, 5G recommendations and enabling technologies, location-aided communication, co-design architectures, and aerial-ground ILAC for 6G.It aims to provide a tutorial overview and discuss interactions across network layers while utilizing infrastructure and radio resources for dual purposes.

1) Self-Localization:

Localization systems differ by where estimation occurs and by how measurements are converted into location. Geometric methods use distances, relative distances, or angles, while RSS-based methods avoid synchronization but generally sacrifice accuracy.

  • Self- and Remote Localization: Self-localization estimates the agent node’s position locally, whereas remote localization sends measurements to a central station for estimation.Remote processing reduces the agent node’s computational burden but centralizes location information.
  • Self- and Remote Localization: Self-localization supports fast local processing, privacy protection, and onboard sensor fusion, but requires substantial device caching and computational capability.These systems are therefore suited to devices with powerful computational components.
  • Self- and Remote Localization: Remote localization is less demanding on agent nodes and preserves locations for all agents in an area, enabling applications such as location-aware communication.Its principal limitation is the privacy and security risk of storing all agent locations on a remote server.
  • Geometric-Based Localization: Two-step localization first extracts RSS, TOA, TDOA, or AOA measurements, then estimates the agent position from those measurements.The generic model is r_n = h(p_n, w) + e_n, where h is nonlinear and e_n represents measurement error.
  • Geometric-Based Localization: RSS estimates distance from signal attenuation without synchronization or LoS dependence, but its localization accuracy is poor, especially in cluttered environments.TOA uses propagation delay for trilateration, while AOA can localize in 2D with two ANs but requires directional antennas or large arrays.

2) Scene Analysis/Fingerprinting-based Localization:

Fingerprinting-based localization replaces geometric inference with pattern matching against a radio map built during offline training. Deterministic and probabilistic methods differ in how they model and match stored fingerprints.

  • Scene Analysis/Fingerprinting-based Localization: Fingerprinting methods use sensor-derived, geotagged signatures to localize users when geometric approaches degrade in complex environments.Possible fingerprints include visual, motion, and signal signatures.
  • Scene Analysis/Fingerprinting-based Localization: RF fingerprinting divides the area into known cells, records RSS fingerprints during offline training, and stores the resulting radio map for online localization.Online measurements are matched against the stored map to estimate the agent’s location.
  • Deterministic Methods: Deterministic fingerprinting selects the cell whose stored fingerprint is closest to the online RSS measurements.NN, KNN, and WKNN are representative deterministic methods; KNN averages the locations of the K closest cells.
  • Probabilistic Methods: Probabilistic fingerprinting estimates location through statistical inference between online measurements and the stored radio map, commonly using MAP or ML estimation.The key quantity is the conditional probability P(sw|z_l).
  • Probabilistic Methods: Parametric probability models require distributional assumptions, whereas non-parametric methods avoid those assumptions but need many time samples per cell for histogram generation.Gaussian, lognormal, and kernel functions are examples of parametric approximations.
  • Scene Analysis/Fingerprinting-based Localization: Fingerprinting is more robust to multipath-induced measurement errors because it transforms localization into offline training and online pattern matching.A location is characterized by its detected signal pattern rather than requiring exact geometric measurements.

3) Proximity-Based Localization:

Proximity-based localization determines an agent’s position from proximity constraints, typically assigning it to the strongest nearby AN. It is simple but intended for systems with modest accuracy requirements.

  • Proximity-Based Localization: Proximity-based localization uses RSS measurements and proximity constraints to identify an agent within a dense grid of ANs.The AN with the strongest RSS is treated as the agent’s location.
  • Proximity-Based Localization: Because proximity estimation provides limited accuracy, it is commonly used in Cell-ID, RFID, and Bluetooth systems with low location-accuracy requirements.It can also reduce the search region before fingerprint pattern matching.
  • Proximity-Based Localization: The surveyed localization approaches are compared by their measurement models, advantages, and disadvantages.

C. Location Estimators

Nonlinear estimators directly minimize a localization cost function and usually achieve high accuracy, but their multimodal objectives may lack guaranteed global solutions and require substantial computation.

  • Location Estimators: Nonlinear estimators directly minimize a cost function constructed from the localization model.They usually provide high localization accuracy, but global optimality may not be guaranteed because the cost functions are often multimodal.

1) Nonlinear Estimators:

The paper contrasts nonlinear and linearized location estimators, relating their assumptions, optimization procedures, and accuracy–complexity trade-offs. Nonlinear methods use model-based cost minimization, whereas linear estimators obtain closed-form solutions through equation linearization.

  • Nonlinear Estimators: NLS minimizes a model-based cost function without requiring assumptions about error statistics.WNLS additionally uses the error covariance, while ML uses the known error probability distribution.
  • Nonlinear Estimators: Under zero-mean Gaussian errors, ML and WNLS have the same performance; ML can attain the CRLB, whereas NLS is simpler when noise information is unavailable.
  • Nonlinear Estimators: Global search can be accurate but time-consuming, while iterative methods require good initialization and may encounter local minima.Newton-Raphson and Gauss-Newton are generally more effective than steepest descent, whereas steepest descent is more stable when the Hessian inverse may not exist.
  • Linear Estimators: Linear estimators such as LLS and WLLS convert nonlinear localization equations into linear forms to obtain closed-form solutions.For 2D TOA, the construction squares the measurement equations and, under sufficiently small noise, eliminates the resulting error terms during linearization.
  • Comparison: The paper provides summaries and comparisons of localization techniques and location estimators in Tables II and III.
  • Linear Estimators: LLS solutions are generally sub-optimal because linearization discards information and perform well mainly when noise is relatively small.WLLS is more generic because it uses measurement mean and covariance information.

D. Main Sources of Error and Mitigation Techniques

Localization performance is limited by estimation biases and measurement errors, motivating analysis of their sources and corresponding mitigation techniques.

  • Main Sources of Error and Mitigation Techniques: The paper identifies estimation biases and measurement errors as fundamental limits on localization performance.It frames robustness improvement around analyzing error sources and their mitigation techniques.
  • Main Sources of Error and Mitigation Techniques: The section introduces three main sources of localization error and discusses mitigation techniques for them.

1) Multi-path Fading:

Multipath, NLoS propagation, and systematic errors degrade localization through unresolved signals, biased measurements, or estimator bias. The paper describes statistical, geometric, calibration, and correction-based mitigation approaches.

  • Multi-path Fading: Multipath fading superimposes signals arriving through different paths, making them unresolvable and complicating signal detection in cluttered environments.
  • NLoS Propagation: NLoS signals weaken the relationship between measurements and link distance by introducing a positive range-estimation bias.
  • NLoS Propagation: NLoS identification can be formulated as statistical detection between LoS and NLoS hypotheses using a differentiating metric.NLoS range measurements tend to have larger variance and non-Gaussian distributions than LoS measurements.
  • NLoS Propagation: When measurements are predominantly NLoS, localization can combine signal measurements with a prior environment map or use measurements from identified scatters.
  • Systematic Error: Systematic errors arise from the localization system, including imperfect measurements, radio miscalibration, clock drift, and clock offset.They bias estimators, remain approximately constant with target location, and cannot be removed by averaging repeated measurements.
  • Systematic Error: Real-time infrastructure calibration, clock-offset correction, and recursive Bayesian methods are described as ways to mitigate systematic errors.

2) Precision:

Precision characterizes the statistical variation of localization accuracy across trials. The paper discusses GDOP, outage probability, CDF-based confidence, and the trade-off between accuracy and complexity.

  • Precision: GDOP measures localization-error variation relative to range-estimation error, with smaller values indicating better localization precision.It also reflects how anchor-node geometry relates to achievable accuracy and can guide anchor placement and selection.
  • Precision: Localization error outage is the probability that localization error exceeds a specified threshold.
  • Precision: The localization-error CDF gives the probability that an estimate meets a predefined accuracy, and higher CDF values indicate better precision when accuracies are equal.
  • Precision: A system with CDF(1.5) = 0.9 achieves 90% precision within 1.5 m, compared with 50% for CDF(1.5) = 0.5.
  • Precision: Localization-system complexity depends on hardware, signal measurement, and estimator computation, creating a trade-off between accuracy and complexity.

4) Coverage and Scalability:

Localization systems trade deployment cost and scalability against achievable performance, while coverage and accuracy vary across infrastructures and environments. Reusing communication networks avoids dedicated infrastructure deployment but generally requires more sophisticated algorithms.

  • Coverage and Scalability: Localization coverage is classified as global, local, or indoor, with performance generally degrading as the agent moves farther from anchor nodes.Coverage denotes the maximum area providing guaranteed accuracy, precision, and latency performance.
  • Localization Infrastructures: Dedicated infrastructures provide high localization performance but incur hardware costs and face scalability limitations.GNSS exemplifies the dedicated-infrastructure approach using specialized reference signals and hardware.
  • Localization Infrastructures: Reusing cellular, WLAN, or other communication infrastructures avoids expensive deployment but typically depends on sophisticated algorithms to improve localization performance.These systems use signals of opportunity to provide localization alongside communication services.
  • Cellular Networks: Cellular localization evolved from 2G time-based methods toward increasingly advanced techniques using both uplink and downlink signals through 4G.Cellular location information also supports commercial services and network optimization.
  • WiFi: WiFi localization is widely available for indoor use but is vulnerable to interference, with typical accuracy of approximately 1 to 5 m and update rates of a few seconds.Decimeter-level accuracy has been achieved in certain scenarios.
  • UWB: UWB supports centimeter-level localization through high temporal resolution, low transmission power, multipath resolution, and obstacle penetration.UWB can support both fingerprinting and geometric-based localization in LoS and NLoS scenarios.

2) SLAM:

Advanced localization methods extend static position estimation to mobile-state tracking, map construction, cooperation among agents, and fusion of heterogeneous measurements. These techniques address unknown anchors, NLoS connectivity constraints, and multipath effects in complex environments.

  • SLAM: SLAM jointly locates unknown anchor nodes and constructs an environmental map while a mobile node follows a predetermined path.The state includes both the mobile node’s tracking variables and anchor locations; EKF is widely used in practice.
  • Cooperative Localization: Cooperative localization lets agent nodes use measurements from both anchors and neighboring agents, improving accuracy and extending coverage in NLoS environments.Noncooperative systems require sufficient direct anchor communication, often demanding denser anchors or longer-range coverage.
  • Data Fusion: Data fusion combines different information types because multipath effects severely limit localization systems that rely on a single measurement type.Heterogeneous wireless technologies motivate hybrid data fusion across cellular, WLAN, RFID, and Bluetooth.
  • 5G and Beyond: Future networks target seamless high-accuracy localization alongside ubiquitous communication and automatic control for Internet-of-Everything services.The cited 5G positioning targets include submeter accuracy for general commercial use cases and specified error and latency bounds for indoor and outdoor deployments.

B. Towards Centimeter Localization for 5G and Beyond

5G technologies such as massive MIMO and mmWave provide communication properties that can also improve localization. Their large antenna arrays, broad bandwidth, directional beams, and sparse multipath enable angular, temporal, and channel-parameter estimation for high-resolution positioning.

  • mmWave Massive MIMO Localization: Massive MIMO improves angle-based localization through highly directional beams and high angular resolution from large antenna arrays.The same arrays also support dense spatial multiplexing and high spectral efficiency for communication.
  • mmWave Massive MIMO Localization: Uniform rectangular arrays steer beams in two dimensions using azimuth and elevation angles.The array response depends on the antenna-grid indices and the signal wavelength-to-element-spacing relationship.
  • mmWave Massive MIMO Localization: mmWave communication uses bandwidth on the order of GHz and carrier frequencies around 30 GHz to 300 GHz, supporting high data rate and low latency.At 60 GHz, bandwidth up to 7 GHz has been standardized for WPANs.
  • mmWave Massive MIMO Localization: mmWave channels typically contain few multipath components, simplifying estimation of path gains, delays, and transmit- and receive-side angles.A beamspace representation makes parameter estimation easier than estimating parameters from the nonlinear antenna-domain channel.
  • mmWave Massive MIMO Localization: MIMO channel-parameter estimation considers azimuth and elevation angles, propagation delay, Doppler shift, and their uncertainties for point and distributed sources.Approaches include subspace-based and compressive-sensing methods.
  • mmWave Massive MIMO Localization: mmWave MIMO localization methods estimate position, velocity, orientation, or environmental maps using hybrid measurements and signal characteristics.Reported results include decimeter-level localization accuracy and position- and rotation-angle estimation bounds.

2) D2D Communication and Cooperative Localization:

Device-centric and dense networks use location information to support cooperative localization and communication across multiple network layers. These approaches can reduce overhead, improve throughput and energy efficiency, and exploit nearby infrastructure or devices for localization.

  • D2D Communication and Cooperative Localization: D2D communication shifts interactions from exclusively BS-mediated links toward direct or relayed device-centric communication.The four categories are device relaying from BS, BS-aided D2D, direct D2D, and device relaying from another device.
  • D2D Communication and Cooperative Localization: Assisting devices can act as pseudo BSs for localizing cell-edge users, while BS-aided links provide measurements from both the BS and another device.Error propagation from assisting devices remains a critical issue for cooperative localization.
  • Localization in UDNs: UDNs are networks with more cells than users; one quantitative definition requires at least 10^3 cells/km2.Small cells can reduce CID-based localization error and increase LoS-link probability because they are closer to users.
  • Location-aware Communication: Location-aware communication uses terminal locations to improve channel estimation, beamforming, resource management, routing, and content delivery.Applications include interference reduction, lower signalling overhead, geographic routing, adaptive streaming, caching, and prefetching.
  • The MAC Layer: Accurate location information supports MAC-layer multicasting, broadcasting, scheduling, and load balancing, with typical accuracy requirements of several meters to tens of meters.These uses address resource sharing and heterogeneous-network traffic imbalance.
  • The MAC Layer: Location-based scheduling can require less feedback and achieve higher total throughput than CSI-based user selection.Location information also supports proactive scheduling and sensor sleep scheduling to reduce transmission or total energy consumption.

D. Localization and Communication Co-design

Localization and communication are commonly designed separately despite sharing radio infrastructure and signals, motivating ILAC co-design. The surveyed architectures allocate or reuse signals, bandwidth, power, and receiver functions while exposing trade-offs between data transmission and localization accuracy.

  • Localization and Communication Co-design: ILAC addresses the relative scarcity of localization-and-communication co-design, whose objectives otherwise differ between localization accuracy and reliable data rate.Existing studies include mmWave beamforming designs that characterize trade-offs between localization efficiency and downlink data rates.
  • Architectures: ILAC architectures combine separate or shared transmitter signals with separate or shared receivers, yielding three architecture categories.A shared receiver can reuse communication channel estimation for location measurement.
  • Separated Signals and Receivers: Bandwidth splitting assigns BL = kBt to localization and BC = (1 − k)Bt to communication over orthogonal frequency bands.Orthogonality avoids interference between the two signals while the fixed total bandwidth creates an allocation trade-off.
  • Separated Signals and Receivers: Increasing dedicated bandwidth improves data rate and decreases the localization CRLB, so bandwidth splitting must match communication and localization QoS requirements.The localization signal’s mean square bandwidth increases with signal bandwidth.
  • Separated Signals and Receivers: Power splitting balances communication and localization because total single-user transmission power is fixed at P = PC+PL.Higher signal power benefits either data transmission or localization accuracy, depending on its allocation.

2) Shared Signal but Separated Receivers:

Shared-signal ILAC requires joint waveform design, while shared receivers can extract location information and decode data from the same signal. These designs extend toward aerial-ground integration for ubiquitous 3D connectivity.

  • Shared Signal but Separated Receivers: Uniform power spectral density is strictly sub-optimal for signal propagation-delay measurement, motivating waveform optimization for shared localization and communication signals.Localization waveform design has received less study than communication waveform design in cellular networks.
  • Shared Signal but Separated Receivers: A shared receiver can combine the LMU and ID functions and reuse the communication CEU for localization when channel parameters are unknown.In multipath environments, location-related estimation is closely related to channel estimation.
  • Aerial-Ground Integration: Aerial-ground integration is a 6G vision for ubiquitous wireless connectivity in 3D space, with UAVs offering mobile and on-demand deployment.UAVs may operate as low-altitude platforms, while floating BSs can operate as high-altitude platforms in the stratosphere.

1) Localization for UAVs:

The paper surveys UAV localization methods and highlights how cellular connectivity, 6D localization, UAV cooperation, and UAV-assisted localization support future 3D wireless networks. Accurate UAV location also benefits communication through radio mapping, CNPC-link management, and 3D beamforming.

  • Conventional and cellular-assisted localization: Vision-based navigation avoids external signals and is cheaper to deploy than INS/GPS, but real-time image processing increases complexity for resource-constrained UAVs.Its accuracy also depends on visual information from the environment.
  • Conventional and cellular-assisted localization: Cellular-connected UAVs can extend localization coverage, reuse communication links as reference signals, and adapt legacy and 5G techniques for aerial localization.Relevant techniques include OTDOA, UTDOA, E-CID, massive MIMO, and mmWave communication.
  • Future UAV localization: 6D localization estimates a UAV’s 3D position and orientation, using massive-MIMO angular resolution and mmWave-supported AOA/AOD estimation.Orientation matters because it affects flying gesture, power consumption, and trajectory design.
  • UAVs assisting network localization: UAV-aided localization uses UAVs as cooperative nodes in FANETs or as additional anchors for terrestrial user equipment.Higher altitude can provide wider coverage and a high probability of line-of-sight links.
  • UAVs assisting network localization: Decimeter-level relative position error and meter-level absolute position accuracy were achieved in one UAV-aided vehicle-localization method using RTOF measurements.Another approach treated a GPS-equipped drone as a mobile anchor and used UWB ranging with trilateration.
  • Localization benefits for communication: Accurate UAV locations support communication through 3D radio-environment maps, proactive CNPC-link management, and beamforming based on aerial-ground azimuth and elevation angles.SNARM uses UAV signal measurements for trajectory optimization and radio-map construction that predicts outage probabilities.
  • Future 3D wireless networks: The envisioned 6G architecture is AI-enabled, heterogeneous, and multi-tier, with distributed and centralized AI supporting signal processing, 3D radio maps, localization selection, and proactive resource management.The architecture spans space, aerial, and terrestrial network segments.

A. Fundamental Performance Analysis and Design for ILAC

ILAC design must address spectrum sharing, cross-layer connectivity, heterogeneous protocols, and radio-resource management in increasingly diverse 3D networks. The paper identifies fundamental analysis, accurate 3D radio-environment maps, and AI-based control as important research directions.

  • Spectrum and waveform design: ILAC must determine whether localization and communication share spectrum or use orthogonal allocations, while jointly designing waveforms for high spectral efficiency.The motivation includes reliable communication and accurate localization for massive IoT connectivity.
  • Cross-layer integration: Multi-tier networks require dynamic maintenance of links across frequency bands, with resource management and signal processing supporting cross-layer information sharing.Channel-estimation units may be reused to extract location information.
  • Protocol integration: Heterogeneous standards and rapidly changing radio environments raise the need for rapid protocol switching without degrading localization and communication performance.Different network layers employ different protocols.
  • 3D radio-environment maps: Proactive resource management depends on accurate radio-environment maps, but existing map modeling mainly targets 2D terrestrial outdoor scenarios.Extending modeling to cluttered 3D environments and relating location accuracy to communication performance remain open issues.
  • AI-enabled ILAC: AI and machine learning can recognize patterns in complex raw radio data and support user-centric networks that autonomously manage resources, functions, and control.The intended objective is sustaining high performance according to real-time user locations.
Loading 2006.01535v1…