PaperlayerView exact revision on arXiv

Source-linked AI summary

5G Embraces Satellites for 6G Ubiquitous IoT: Basic Models for Integrated Satellite Terrestrial Networks

Xinran Fang, Wei Feng, Te Wei, Yunfei Chen, Ning Ge, Cheng-Xiang Wang

arXiv:2011.03182v2cs.IT

TL;DR

HSTNs must reconcile the complementary coverage and performance characteristics of satellite and terrestrial communications without reducing their coupled behavior to separate links or an intractable whole-network model. The paper addresses this gap by surveying three tractable cooperative models—X, L, and V—and identifying directions for future HSTNs. It concludes with a mesoscopic framework for understanding heterogeneous HSTNs and open issues for agile, smart, and secure 6G ubiquitous IoT.

  • Problem

    Differences in channel fading, delay, mobility, and coverage prevent simple HSTN decomposition, while whole-network treatment is too complex.

  • Method

    The paper models HSTNs as combinations of three basic cooperative models and surveys technologies, research problems, and solutions for each.

  • Results

    The paper provides a mesoscopic framework connecting micro link analysis and macro network evaluation through models X, L, and V.

  • Takeaways & Limitations

    The models support future directions toward cell-free, hierarchical, decoupled HSTNs for 6G ubiquitous IoT.

Abstract

from arXiv · show

Terrestrial communication networks mainly focus on users in urban areas but have poor coverage performance in harsh environments, such as mountains, deserts, and oceans. Satellites can be exploited to extend the coverage of terrestrial fifth-generation (5G) networks. However, satellites are restricted by their high latency and relatively low data rate. Consequently, the integration of terrestrial and satellite components has been widely studied, to take advantage of both sides and enable seamless broadband coverage. Due to the significant differences between satellite communications (SatComs) and terrestrial communications (TerComs) in terms of channel fading, transmission delay, mobility, and coverage performance, the establishment of an efficient hybrid satellite-terrestrial network (HSTN) still faces many challenges. In general, it is difficult to decompose a HSTN into a sum of separate satellite and terrestrial links due to the complicated coupling relationships therein. To uncover the complete picture of HSTNs, we regard the HSTN as a combination of basic cooperative models that contain the main traits of satellite-terrestrial integration but are much simpler and thus more tractable than the large-scale heterogeneous HSTNs. In particular, we present three basic cooperative models, i.e., model X, model L, and model V, and provide a survey of the state-of-the-art technologies for each of them. We discuss future research directions towards establishing a cell-free, hierarchical, decoupled HSTN. We also outline open issues to envision an agile, smart, and secure HSTN for the sixth-generation (6G) ubiquitous Internet of Things (IoT).

I. INTRODUCTION

The paper frames HSTNs as a response to expanding IoT connectivity needs and the limitations of separate satellite and terrestrial networks. It proposes three tractable cooperative models to bridge detailed link analysis and complex network-wide evaluation.

  • IoT applications are expanding across transportation, monitoring, agriculture, autonomous driving, health care, and disaster recovery.
  • SatComs face high path loss, interference from overlapping satellite beams, and high broadband service costs.
  • HSTNs are difficult to analyze because major SatCom and TerCom differences prevent simple decomposition into separate links.
  • The paper presents models X, L, and V as simpler cooperative abstractions of large-scale heterogeneous HSTNs.
  • For each model, the paper surveys state-of-the-art technologies, discusses performance, resource management, networking, and security, and identifies future directions.

II. SYSTEM OVERVIEW

An HSTN integrates satellite, aerial, and terrestrial domains to strengthen coverage across urban, suburban, and remote environments. Its heterogeneous channels, delays, mobility, coverage, and resource constraints make efficient design challenging.

  • A. Application Scenarios: An HSTN combines terrestrial infrastructure with satellites, gateways, HAPs, aerial platforms, and satellite or dual-mode terminals.
  • A. Application Scenarios: HMTs can use terrestrial networks within coverage and satellites when terrestrial service is unavailable.
  • A. Application Scenarios: Satellites strengthen incomplete terrestrial coverage in remote regions, while LEO networks, high-altitude relays, and UAVs support offloading or complementary coverage.
  • B. Differences Between SatComs and TerComs: SatComs differ from TerComs in propagation loss, delay, scattering, attenuation, mobility, and coverage performance.
  • B. Differences Between SatComs and TerComs: Serving diverse QoS demands under limited spectrum and power creates complex, dynamic interference through resource reuse.
  • B. Differences Between SatComs and TerComs: Basic cooperative models provide a tractable alternative to both separate-link decomposition and treating the entire heterogeneous network as one unit.

III. BASIC COOPERATIVE MODELS

The paper represents heterogeneous HSTNs through three minimal cooperative models that capture satellite-terrestrial interactions at a mesoscopic level. These abstractions connect micro link analysis with macro network evaluation.

  • HSTNs can be represented as combinations of three basic cooperative models using the fewest wireless links.
  • Each basic model contains one satellite link, one terrestrial link, one satellite, one satellite user, one base station, and one terrestrial user.
  • Model X: Model X uses separate satellite and terrestrial communications that share the same spectrum resources.
  • Model L: Model L uses a relay combining base-station and satellite-terminal functions between a satellite and its user.
  • Model V: Model V has a satellite and terrestrial base station jointly serve one common hybrid mobile terminal.
  • Studying the models can provide a mesoscopic view that helps characterize arbitrary HSTNs between micro and macro analysis.

A. Model X

Model X studies spectrum sharing between a satellite link and a terrestrial link, where their mutual interference has diverse and complicated patterns because of differing coverage and communication characteristics.

  • Model X shares spectrum between a satellite communicating with an ST and a TBS communicating with a GMT.
  • The model’s key technical challenge is mutual interference between the satellite and terrestrial links.
  • Interference patterns differ from those in homogeneous networks because satellite-terrestrial differences create diverse and complicated interactions.
  • In urban areas, STs mainly experience interference from adjacent TBSs, while satellite interference from GMTs is almost negligible compared with TBS interference.
  • For TBSs and GMTs, interference from neighboring STs is usually stronger than interference from satellites.
  • The section surveys model X studies in system performance, resource management, and networking.

1) Performance Analysis:

Research on model X covers performance analysis and resource management under satellite-terrestrial interference, while practical channel, CSI, delay, and hardware constraints remain important gaps.

  • Performance Analysis: Studies analyze model X using capacity, ergodic capacity, and outage probability, including theoretical, experimental, fading, and interference-based evaluations.
  • Performance Analysis: Shadowed-Rician fading is the most widely used satellite-ground channel model, while comprehensive Nakagami-m evaluations remain open.
  • Performance Analysis: Imperfect CSI makes accurate interference evaluation difficult, motivating more practical performance analyses.
  • Resource Management: Resource-management research addresses power allocation, spectrum sharing, carrier allocation, and beamforming to influence performance under shared spectrum.
  • Resource Management: Power-allocation objectives include capacity, QoS, delay-limited capacity, outage probability, energy efficiency, rate, and interference mitigation.
  • Resource Management: Spectrum-sharing and beamforming methods target spectrum scarcity, mutual interference, high path loss, spectrum efficiency, and secure transmission.

3) Networking Issue:

Networking studies address the integration of satellite and terrestrial subnetworks through heterogeneous architectures, programmability, virtualization, and resilience mechanisms.

  • Networking Issue: Extending model X to satellite and terrestrial subnetworks creates networking challenges because the two networks have substantially different characteristics.
  • Networking Issue: Proposed architectures jointly exploit 5G, Wi-Fi, Bluetooth, LoRa, and ground, aerial, space, and undersea components for HSTN and IoT applications.
  • Networking Issue: SDN and NFV support programmability, openness, virtualization, virtual spectrum allocation, dynamic traffic offloading, and end-to-end delay analysis.
  • Networking Issue: An extension of model X produces a three-dimensional cell-free HSTN rather than a generally regular cellular architecture.
  • Networking Issue: SDN-enabled HSTNs have been studied for elastic resilience, including controller reachability, failure detection, and recovery.
  • Networking Issue: Future networking studies should account for limited cost, nonlinear hardware, and imperfect prior knowledge.

4) Future Directions:

Future directions target cell-free, dynamically decoupled HSTNs while accounting for relay-assisted model L configurations and unresolved analytical challenges.

  • Extending model X with multiple satellite and terrestrial links leads to a three-dimensional cell-free architecture instead of geographically separated cellular cells.
  • Because the cell-free architecture is undecomposable and difficult to analyze, dynamic decoupling on radio maps is proposed as one possible solution.
  • Radio maps may characterize large-scale interference relationships, but the basic theory and methods for this approach remain unknown.
  • Model L typically uses one satellite, one relay, and one destination user, with TBSs or UAVs serving as relays.

1) Performance Analysis:

Research on model L evaluates reliability, capacity, effective capacity, and outage-related performance across amplify-and-forward and decode-and-forward relaying, with extensions using multiple relays, hops, antennas, and users.

  • Performance Analysis: Model L performance studies analyze symbol error rate, average symbol error rate, capacity, effective capacity, and outage probability.Existing work is organized mainly by amplify-and-forward and decode-and-forward relaying modes.
  • Performance Analysis: Most studies examine one-way relaying, with satellite-to-relay transmission followed by relay-to-destination forwarding.Analyses include AF relays over non-identical fading channels, Alamouti HSTNs, and MPSK average symbol error rates.
  • Performance Analysis: Enhanced model L studies add multi-antenna satellites and users, optimistic scheduling, NOMA, multiple relays, and multi-hop relays.These extensions target multi-user performance and spectrum efficiency through additional spatial or relay resources.
  • Performance Analysis: Adding antennas and relays can improve performance, but existing studies focus on gains without modeling cost or optimizing benefit-to-cost ratio.The paper identifies cost modeling and BCR optimization as necessary for practical deployment.
  • Performance Analysis: Decode-and-forward generally outperforms amplify-and-forward because of decoding gain.Studies also examine hardware impairments, NOMA, multiple UAV relays, and opportunistic relay selection.

2) Resource Management:

Model L resource-management research addresses asymmetric roundtrip time through relay selection, power allocation, beamforming, secure transmission, energy efficiency, and delay-aware allocation.

  • Resource Management: Asymmetric roundtrip time makes model L resource management different from conventional scenarios.Research includes joint relay selection and power allocation, relay beamforming, and secure-rate maximization.
  • Resource Management: Resource-allocation studies optimize beamforming, direct-versus-relayed transmission, energy efficiency, remote-user information forwarding, and backhaul delay requirements.These methods coordinate resources across satellite, relay, and terrestrial links.
  • Resource Management: Open security work derives secrecy capacity with eavesdroppers and AF relays and designs beamforming to maximize HSTN secrecy rate.Satellite links’ openness and long transmission distances create security concerns addressed through relaying and beamforming.
  • Resource Management: Model L research still underexplores transmission-delay differences, processing delay, satellite mobility, handover time, adaptive relay processing, and system cost.Matching SatCom and TerCom delays and optimizing benefit-to-cost ratio remain future directions.
  • Resource Management: An arm-hand-like extension of model L combines satellite large-scale coverage with relay beams for high-precision user targeting.Relays can be deployed on mobile platforms such as cars, vessels, or airplanes to create elastic hierarchical coverage.

C. Model V

Model V uses a hybrid terminal connected to one satellite and one terrestrial base station, enabling terrestrial, satellite, or coordinated access across changing coverage conditions.

  • C. Model V: Model V consists of one satellite, one BS, and one hybrid mobile terminal.The terminal can connect to terrestrial or satellite systems depending on availability and service conditions.
  • C. Model V: In urban areas the terminal mainly uses terrestrial broadband, while in mountainous, disaster, desert, and marine areas it turns to satellites.Satellite-terrestrial coordination can also assign low-rate wide-area signaling to satellites and high-speed local data to TBS pencil beams.
  • C. Model V: Multi-diversity reception in model V uses maximal-ratio combining or selective combining to compensate for satellite attenuation.MRC sums incoming carrier-to-noise ratios, whereas SC selects the best signal.
  • C. Model V: Model V requires careful access-point selection and radio-resource allocation across satellite and terrestrial links.Research considers multi-radio access, scheduling, beamforming, user scheduling, routing, and post-disaster communication.
  • C. Model V: Mobility creates horizontal handovers between satellites or TBSs and vertical handovers between satellite and terrestrial systems.Bidirectional satellite-user mobility increases handover uncertainty and can produce long delays during cross-domain switching.

4) Future Directions:

Future HSTN work targets lower-overhead cooperation, coordinated control and communication, and intelligent integration technologies while addressing latency and evolving standardization requirements.

  • 4) Future Directions: Model V cooperation may incur high inter-system communication complexity and overhead, requiring protocol transformation and rate matching.Future work should study low-overhead information transfer and cooperative interaction across mismatched satellite and terrestrial systems.
  • 4) Future Directions: Model V extensions can separate wide-area satellite signaling from terrestrial pencil-beam data delivery to target users efficiently.The satellite control subsystem can identify demand in TBS blind areas, while the TBS focuses resources on corresponding users.
  • 4) Future Directions: Latency differences between satellite and terrestrial systems make timely control-communication interaction challenging.Extrinsic information such as shipping lanes and historical user behavior are proposed for on-demand and prediction-driven responses, but remain an open issue.
  • 4) Future Directions: 3GPP completed NTN standardization in Release 16, while new 5G NR normative solutions and longer-term studies continue in Releases 17–19.The paper places future HSTN development within an ongoing standardization process.
  • 4) Future Directions: Artificial intelligence, MEC, and blockchain are identified as frontier technologies for smart integration of satellite and terrestrial communications.The paper presents these technologies as potential contributors to future HSTN development.

A. BCR Maximization for Basic Cooperative Models

The paper frames HSTN development around evaluating gains against costs and adapting model combinations to changing conditions. It identifies learning, integrated caching–computing–communication, and blockchain as technologies for future HSTN management.

  • BCR evaluation: HSTN design should evaluate benefit-to-cost ratios (BCRs) using holistic cost models and mathematical characterization of multidimensional gains.The paper proposes establishing a BCR optimization framework for each basic cooperative model.
  • Intelligent orchestration: Intelligent orchestration dynamically changes the combination of basic cooperative models according to service requirements, environmental information, and network status.A cyber agent is proposed to make orchestration decisions and learn from historical behaviors.
  • Learning technologies: Deep reinforcement learning can address hard-to-model HSTN problems, while federated learning requires redesign for differing satellite and terrestrial network conditions.The paper emphasizes that SatCom–TerCom differences constrain direct use of conventional learning methods.
  • Integrated services: Smart caching and mobile edge computing can integrate caching, computing, and communication, but their coupled complexity calls for systematic and economic methods.Smart pricing is given as an example of an economic methodology.
  • Resource allocation: Blockchain may support an open ecology for HSTN resource allocation while reducing management complexity and potentially improving efficiency and security.Satellite broadcasting may also improve traditional blockchain efficiency, including transactions per second.
  • Mesoscopic modeling: The three basic cooperative models provide a mesoscopic perspective connecting micro link analysis with macro network evaluation of HSTNs.This perspective is intended to make large-scale heterogeneous HSTNs more tractable to understand.
Loading 2011.03182v2…