The Potential of LEO Satellites in 6G Space-Air-Ground Enabled Access Networks

Ziye Jia, , Chao Dong, , Kun Guo, , and
Qihui Wu Ziye Jia is with the College of Electronic and Information Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China, and also with the State Key Laboratory of ISN, Xidian University, Xi’an 710071, China (e-mail: jiaziye@nuaa.edu.cn). Chao Dong and Qhui Wu are with the College of Electronic and Information Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China (e-mail: wuqihui@nuaa.edu.cn, dch@nuaa.edu.cn). Kun Guo is with the East China Normal University, Shanghai 200241, China (e-mail: kguo@cee.ecnu.edu.cn).

Abstract

Space-air-ground integrated networks (SAGINs) help enhance the service performance in the sixth generation communication system. SAGIN is basically composed of satellites, aerial vehicles, ground facilities, as well as multiple terrestrial users. Therein, the low earth orbit (LEO) satellites are popular in recent years due to the low cost of development and launch, global coverage and delay-enabled services. Moreover, LEO satellites can support various applications, e.g., direct access, relay, caching and computation. In this work, we firstly provide the preliminaries and framework of SAGIN, in which the characteristics of LEO satellites, high altitude platforms, as well as unmanned aerial vehicles are analyzed. Then, the roles and potentials of LEO satellite in SAGIN are analyzed for access services. A couple of advanced techniques such as multi-access edge computing (MEC) and network function virtualization are introduced to enhance the LEO-based access service abilities as hierarchical MEC and network slicing in SAGIN. In addition, corresponding use cases are provided to verify the propositions. Besides, we also discuss the open issues and promising directions in LEO-enabled SAGIN access services for the future research.

Index Terms:

Space-air-ground integrated network (SAGIN), low earth orbit (LEO) satellite, multi-access edge computing (MEC), network slicing.

I Introduction

I-A Background and Motivation

The sixth generation communication system (6G) related researches have gained great interests from both industries and academics, among which the space-air-ground integrated network (SAGIN) is a promising technique to extend the range of global services and worldwide Internet access [1, 2, 3]. Specifically, with the explosive growth of satellite businesses from various individuals and institutions, such as satellite navigation, emergency rescue, and deep space exploration, there exists significant demands to enhance the performance of traditional communication networks [4], and SAGIN related techniques are promising to satisfy such requirements. Besides, in order to enable the region coverage and access quality, with the limitation of ground infrastructures, SAGIN enabled access and global coverage are significant.

In particular, SAGIN is typically composed of three layers according to the altitude, i.e., space satellites, aerial vehicles, and terrestrial facilities, as the illustration in Fig. 1. Specifically, in the space, satellites are basically divided into three types based on different heights, i.e., geosynchronous earth orbit (GEO) satellites, medium earth orbit (MEO) satellites, and low earth orbit (LEO) satellites [5, 6]. The high altitude platforms (HAPs) and unmanned air vehicles (UAVs) are popular facilities in the air [7, 8]. HAPs are characterized for the stable and large coverage, while UAVs are specialized in flexible services [9]. Besides, there exists heterogenous resources such as sensors, transmitters, computation modules, caches, etc. Among which, LEO satellites play prominent roles in terms of networking and constellations, and we mostly focus on elaborating the roles and potentials of LEO satellites in the SAGIN access networks.

However, there are multiple heterogenous users such as remote Internet of things (IoT), maritime voyages, and air transportation. Furthermore, the number of such users has a sharp increment in recent years. It is challenging to satisfy the ever-increasing requirements. However, the number of platforms in SAGIN, as well as the corresponding loading capacity such as hardwares, are limited. Besides, the resources in SAGIN are dynamic and the operational modes of various platforms are quite diverse. For example, LEO satellites in different orbits have different cycle periods, the movement of UAV is flexible but unpredictable, while HAPs are quasi-static. Also, the scarce intermittent communication resources caused by the periodic motion of satellites aggravate the resource competition. Consequently, how to figure out such challenges, i.e., heterogeneous platforms, various demands with diverse quality of service (QoS) requirements, and limited resources, are significant issues.

As above, an elastic and scalable SAGIN should be well designed to satisfy the various demands. Besides, the potential of LEO satellites should be explored due to the increasing number of various LEO satellite constellation, such as Starlink [10]. The advanced in-network computing paradigms such as multi-access edge computing (MEC), network function virtualization (NFV), and software defined network (SDN), are introduced into the SAGIN for efficient and resilient resource providers [11]. SDN can assist manage SAGIN by the pattern of separating the control plane and data plane. NFV helps to provide multiple services for different users via decoupling the virtual network functions (VNFs) from physical platforms. In addition, a serious of VNFs for one task form a service function chain (SFC). The MEC paradigm can be utilized to help the remote users in SAGIN for computation task offloading [12]. In short, MEC, NFV, and SDN are prospective modes for flexible resource provider, and can help deal with multiple tasks with heterogeneous platforms and limited resources in SAGIN.

Refer to caption — Figure 1: A view of SAGIN scenario.

I-B State of the Art

The SAGIN as well as LEO satellites, are widely studied recently. For example, in [4], the authors focus on the data offloading problem in the space-air-ground network to guarantee the balance between energy consumption and mean time cost. [13] investigates the LEO satellite access networks by integrating with terrestrial network to realize seamless global communication services. In [14], the authors investigate the distributed mobility management framework of the space-terrestrial networks, via reconfiguration of mobility management functions, to improve the handover decision efficiency. In [15], the authors analyze the communication system of LEO satellites to investigate the application utilizing stochastic geometry. [10] investigates the LEO satellite constellation to improve the Internet connections for remote areas. [16] presents a multi-layer management structure composed of MEO and LEO satellites to achieve efficient mobility and resource control.

As for the MEC and NFV in SAGIN, [6] designs a LEO satellite enabled heterogeneous MEC framework, as well as an offloading scheme to improve the computational performance. In [11], a software defined LEO satellite framework as well as VNF deployment model are proposed, and efficient deployment algorithms are designed. [17] proposes a reconfigurable intelligent surfaces based MEC framework of space information networks to improve both the communication and computing capabilities. In [18], the authors design a LEO satellite aided edge computing platform to guarantee the computing continuum. [19] proposes a framework of orbital edge computing with LEO satellite constellations, to satisfy the growing demand of multiple applications. [20] analyzes a serious of possible ways to combine machine learning and satellite networks to provide satellite based computing.

TABLE I: Platforms in the space and air.

Classification	Space	Air
Classification	Satellite	HAP/Airship	UAV
Height	$\geq$ 160km	17km-22 km	600m-18km
Time duration	Years	Months to years	Minutes to hours
Coverage	Large and periodic	Medium and fixed	Small and dynamic
Energy provider	Solar panel	Solar panel or battery	Battery
Controller	Ground	Satellite or ground	Ground
Cost	High	Medium	Low
Applications	Observation Communication Navigation Astronomy Meteorology …	Real-time monitoring Communication relay
Emergency recovery
Rocket launch platform …	Sensing Communication
Surveillance Farming Logistics …
Ref	[11, 21, 14]	[7, 22]	[8, 23]

However, from the perspective of LEO satellites, the multi-tier MEC for SAGIN, as well as the network slicing related researches have not been thoroughly investigated. Hence, in this work, we explore the potential of LEO satellite based access patterns in SAGIN, especially the heterogenous MEC and NFV-based network slicing.

I-C Contributions

In this work, we provide the preliminaries and current development of SAGIN, and analyze the access patterns from the perspective of LEO satellites. Besides, advanced techniques such as MEC and NFV are introduced in the LEO satellites based access networks to promote the network ability and enhance the service quality, in terms of served users’ number, service diversity, resource flexibility, etc. Both users and resource providers benefit from the new-type access technologies. In order to differentiate the general single-layer mobile computing service, we further present the heterogenous MEC framework as well as the SFC deployment technique. To clearly elaborate the performance, use cases are also provided. Moreover, we provide preliminary analyses on the open issues and possible directions for LEO satellite based access in SAGIN. Note that in this work, we mainly consider the remote users and non-terrestrial resources, instead of the abundant urban wireless communication facilities. The contributions of this work are summarized as follows:

•

The preliminaries of SAGIN are detailed, in which the LEO satellites based access patterns are presented, including the employment of advanced techniques such as MEC, NFV, caching, etc.
•

Based on LEO satellites, the hierarchical MEC as well as SFC implementation based network slicing for SAGIN are investigated. Besides, corresponding use cases are provided to validate the proposed schemes.
•

The LEO satellites based access possibilities in SAGIN as well as open challenges are analyzed, also with promising directions for the future researches.

The rest of this work is organized as follows. We firstly provide an overview of SAGIN in Section II, including the compositions, resources, demands, applications, etc. In Section III, the role and potential of LEO satellites based access in SAGIN are elaborated, as well as the advances of in-network computing paradigms MEC and NFV based resource management paradigms, including validated use cases. The open issues and possible directions and are presented in Section IV. Finally, we draw the conclusion in Section V.

II Basics of SAGIN

In this section, we provide the basics of SAGIN including LEO satellites. Specifically, as illustrated in Fig. 1, SAGIN is generally composed by the hierarchical satellites in the space, multiple vehicles in the air, and basic infrastructures on the ground. As for the multiple satellites in the space segment, the non-LEO satellites include GEO and MEO satellites. For example, the developed GEO/MEO satellites include MicroGEO [14], Boeing, O3B, etc. The air vehicles include airships, HAPs, aircrafts, and UAVs, on the basis of different heights and functions. Besides, most users come from ground, such as emergency areas, ocean, deserts, and remote areas without services of ground base stations. Note that with the development of deep space exploration, e.g. lunar exploration and Mars exploration, the deep space probes are also potential users in SAGIN, since the explored data should be transmitted back to earth via multi-layer satellites. In addition, agents such as satellites, HAPs, and UAVs, act as resources in some cases, and as users in a couple of scenarios. Note that the meta resources include sensors, transceivers, computations, caching, etc., equipped on various platforms, to support multiple applications, such as deep space exploration, air transportation, ocean service, emergency rescue, remote area users, and navigation augmentation [24, 25]. Hence, the resource management and efficient allocation are significant issues, to leverage limited resources to satisfy heterogeneous demands.

The properties of satellites, HAPs, and UAVs are provided in Table I. Specifically, from the view of height, satellites in the space operate above 160km, HAPs or airships in the air implement among 17km-22km, and UAVs in the air have a flight height among 600m-18km, depending on the detailed demands and platform property. Besides, satellites can work during many years, HAPs can last for months to years, and UAVs can continue working as minutes to hours. The coverage of satellites is large and periodic, HAPs can cover an medium and fixed area, while UAVs can flexibly cover a small area. From the perspective of energy provider, satellites utilize the solar panel, HAPs leverage the solar panels as well as battery to store energy at night, and lithium batteries are applied in UAVs. In addition, the controllers for satellites are from ground, HAPs can be controlled by the ground center or satellites, and the existing UAVs are operated by the ground controller. The typical applications of satellites include observation, communication, navigation, astronomy, meteorology, etc. HAPs can support applications such as real-time monitoring, communication relay, emergency recovery, and rocket launch platforms. UAVs are applied in the areas of sensing, communication, surveillance, farming and logistics.

TABLE II: Satellite preliminaries

Satellites	Orbit altitude	Orbit period	Launch cost	Round-trip delay	Network complexity	User switching
LEO	500km-2000km	1.5h-2h	Low	30-50ms	High	Frequent
MEO	2000km-20,000km	2h-24h	Medium	125ms-250ms	Medium	Medium
GEO	20,000km-36,800km	24h	High	400ms-600ms	Low	Not necessary

III LEO Satellite based Access

In this section, the characteristics of LEO satellite based access patterns in SAGIN are analyzed firstly, and further we introduce the network advances for enhancement, i.e. the hierarchical MEC as well as the network slicing techniques. Besides, correlated use cases are provided to verify the corresponding technique implementation.

III-A Properties of LEO Satellites

LEO satellites attract more attentions from the world due to the networking cooperation via inter-satellite links and mega-constellation, such as Starlink of SpaceX and Kuiper of Amazon, to provide abundant applications for increasing demands from remote IoT, maritime voyage, etc. To make it clear, in Table II, by revisiting the properties of multiple satellites, the potential of LEO satellites is discussed. In particular, the height of LEO satellites ranges from 500km to 2000km, the height of MEO satellites is among 2000km to 20,000km, and GEO satellites operate between 20,000km and 36,800km. Besides, the height larger than 36,800km is deemed as outer space. The orbit periods of LEO, MEO, and GEO satellites are 1.5h-2h, 2h-24h, and 24h, respectively. As for the launch cost, LEO satellites are the lowest. In addition, the round-trip delay of LEO satellites is 30ms-50ms, while it is 125ms-250ms and 400ms-600ms for the MEO and GEO satellites, respectively. Hence, compared with MEO and GEO satellites, LEO satellites can provide lower-delay services for the terrestrial users. Besides, LEO satellite constellations dramatically improve the global coverage, including the polar areas. However, the network complexity of LEO satellites is high, since many LEO satellites are networking, and serious in the mega-constellation, while the network complexity of GEO satellite is low since it can operate dependently to cover 1/3 globe. Indeed, users suffer from the frequent handover within the coverage of LEO satellites, while it is not even necessary to switch due to the large coverage and quasi-static property of GEO satellites.

From the perspective of LEO satellites, users may come from deep space, space, air, and ground (including ocean). Moreover, LEO satellites act as multi-hop relays or servers. In short, the properties and advantages of LEO satellites reveal why most countries in the world pay close attentions to LEO network constructions. Accordingly, the challenges such as network complexity and frequent switch as discussed above should be addressed.

III-B Advances for Enhancement

The emerged advances such as MEC, and NFV based network slicing techniques, enable LEO satellites access to flexibly accommodate more users, and efficiently leverage the heterogenous resources. Fig. 2 provides a couple of access examples, from the perspective of LEO satellites, such as the LEO based direct access, LEO assisted UAV access, LEO assisted HAP access, etc. Besides, the advanced techniques such as MEC, cache, and NFV are also applied in the LEO satellites enabled access network to enhance performance. Note that other than the direct offloading from terrestrial users to LEO satellites in the example references [4, 6], the LEO satellites enabled multi-layer MEC mechanism is more applicable. Besides, the LEO based slicing in SAGIN is prospective, instead of independent LEO satellites based VNF deployment in Fig. 2. More specifically, MEC can be deemed as the node-level resource virtualization technique, while SFC implementation is the network-level virtualization technique, i.e., network slicing. Accordingly, the LEO satellite enabled hierarchical MEC as well as the network slicing are elaborated in the following.

III-B1 LEO Satellites Enabled Hierarchical MEC in SAGIN

The MEC paradigm can provide efficient and effective mechanism to assist terrestrial users to deal with the computation demands. In SAGIN, satellites, HAPs, UAVs, as well as ground users, are equipped with different computation capability according to limited load capacity. The multi-layer MEC enables the computation continuum. As illustrated in Fig. 3, the terrestrial terminal (user 1) is equipped with limited computation server capability, so the task is partly computed locally by user 1, and partly offloaded to the LEO satellite for online computing. As for terrestrial user 2 and user 3, the task data are offloaded to the nearby UAVs, and the computation task of user 2 is partially completed by the UAV. However, due to the limited loading capacity of UAVs, a portion of data of user 2 is offloaded to the LEO satellite for further computation. The data from user 3 is offloaded the satellite relayed by a UAV, while the data from user 4 is transmitted by a UAV to the HAP for computation. Note that the data from user 2 and user 3 share the computation resources of the same LEO satellite.

In particular, according to [18, 23], there exists the following computation offloading modes: binary and partial. The binary mode means the data is fully offloaded or completed locally, as the data offloading mode of user 3 and user 4 in Fig. 3. The partial mode means the cooperation of different platforms to complete the same computation task, such as the service procedures of user 1 and user 2 in Fig. 3. Indeed, the detailed offloading decision policy should be determined according to optimization objective and considered metrics such as limited energy supply, communication resource limitation, delay QoS, etc. The general methods such as game theory, convex optimization, and machine learning are available.

Numerical example

We explore the performance of different MEC paradigms in Fig. 4. Specifically, the scenario is set with four LEO satellites with orbit height of 1000km, four UAVs are uniformly distributed in a fixed area with altitude of 2km, and 20-120 users are within the coverage. The hierarchical MEC is implemented by the cooperation of UAVs and LEO satellites, while the compared methods are single-layer LEO satellites and single-layer UAVs based MEC. From Fig. 4, it is observed that with respect to the number of served users, the hierarchical MEC is outstanding compared to the single-layer MEC, and the result is in accordance with intuitive understanding. Besides, the LEO-based MEC performs better that the UAV-based MEC due to the larger coverage and stronger computing load capacity of LEO satellites.

III-B2 LEO Satellites Based Network Slicing in SAGIN

Recently, due to the advantages such as programmability and virtualization introduced by NFV and SDN, the related network slicing techniques are introduced into satellite networks, aiming at cost-effective resource utilization and high QoS performance, as well as lowering the capital and operation expenditures [11]. The basics depend on the advanced reconfigurable resources of both software and hardware. SDN is equipped with the feature of data and control separation, and operating in a centralized management mode. The specific implementation is utilizing NFV technique to provide multiple services for different users via decoupling the VNFs from physical platforms. In addition, a series of VNFs for one task form a SFC. Moreover, the network slicing can also be expanded in the SAGIN, instead of only satellite networks. Such characteristics facilitate an elastic SAGIN, and the SFC deployment in the SAGIN is a significant mode for efficient resource implementation.

For clarity, in Fig. 5, a case of network slicing scenario via SFC deployment in SAGIN is presented. In particular, LEO satellites in the space and the aerial vehicles can provide flexible services leveraging SDN and NFV techniques, and the network slicing can be realized via SFC deployment in the SAGIN. In detail, there exists three different SFC requirements in Fig. 5: the SFC of user 1 includes only one VNF and it is deployed on a LEO satellite, and the UAV serves as a relay. The SFC corresponding to user 2 is composed of three VNFs, which are sequently deployed on a UAV, a HAP, and a LEO satellite, respectively. The SFC from the ocean user 3 includes two VNFs, which are respectively mapped on a UAV and a LEO satellite. Note that the SFCs of user 1 and user 2 share the resources of the same LEO satellite.

The network slicing in SAGIN improves the resource efficiency, for example, increasing stable coverage, reducing latency, etc. Fundamentally, the related problem should consider SFC deployment optimization, multiple resource limitation, the coupled restriction between VNF mapping and routing selection, as well as time horizon related scheduling should be handled.

Numerical example

To clearly verify the performance of LEO satellites enabled network slicing in SAGIN, the numerical results are conducted in a simple use case and Fig. 6 shows the results. Specifically, the scenario is set within a small scale Walker constellation composed of sixteen LEO satellites with orbit height of 1000km, six UAVs are uniformly distributed in a fixed area with altitude of 2km, and 10-100 tasks with SFC requirements. In Fig. 6, different paradigms including flexible deployment, random deployment, and fixed deployment are compared. It is obviously that the flexible SFC deployment performs best, while the random deployment leads to multiple conflicts and resource inefficiency, which typically results in SFC failure. As for the fixed deployment, it means NFV is not supported in such a strategy, and VNFs should be deployed on specified platforms, so the SFC request is liable to fail if there is no active communication links.

IV Open Issues and Directions

TABLE III: Open challenges and possible directions of LEO-access in SAGIN.

Open challenges	Opportunities	Directions
Large number of LEO satellites of mega-constellation	Co-design of various LEO satellite constellations	Interoperability among different constellations via protocol design
High dynamics and frequent users’ handover	Cooperation of multiple platforms for continuum	Mobility management, distributed manage system
On-board processing and control	Intelligent platform	Online learning, distributed control
Information safety	Privacy protection	Differential privacy, blockchain

Although LEO satellites open up significant potentials in the SAGIN access networks, there still exist a couple of open issues, such as large number of LEO satellites in the mega-constellation, high dynamics leading to frequent handover, on-board processing and control, information safety, etc. For clarity, we summarize such issues in Table III. To tackle these challenges, we provide the following possible directions and tips for future researches.

•

As for the large number of LEO satellites in the mega-constellation, the opportunities rely on the co-design of various LEO satellite constellations, and the future directions may advocate increasing the interoperability among different LEO satellites from different companies or countries via protocol design.
•

In order to deal with the high dynamics of LEO satellites, and the related frequent handover issues of users, the cooperation of multiple platforms can guarantee continuum to some extent. Besides, the future directions may depend on the mobility management strategy design, efficient handover decisions, and developing the distributed manage system.
•

To figure out the challenge of on-board processing and control, typically the intelligent and machine learning based techniques can be employed. In particular, the online learning mechanism design as well as the distributed control paradigms may be promising approaches for future directions.
•

Due to the global service and openness of SAGIN, especially LEO satellite networks, the information safety correspondingly becomes intractable issues. Associated with the opportunity of privacy protection, the emerged popular techniques such as differential privacy as well as blockchain can be considered as future directions for SAGIN safety.

Essentially, the open challenges of the LEO satellites access in SAGIN are not only in terms of the discussed issues as above, the orbit scarcity, frequency competition, and resource ability limitation are all issues should be focused, and available strategies should be devised.

V Conclusions

In this work, we have reviewed the basics and details of LEO satellites based access modes and techniques in SAGIN. Firstly we have elaborated the SAGIN preliminaries and current development, as well as the possible applications. Then, the potentials of LEO access patterns, including the direct access, multi-hop, relay, caching, MEC, and NFV implementation have been elaborated. Besides, based on the recent advanced techniques MEC and NFV, the multi-layer MEC as well as the SFC deployment network slicing in SAGIN have been detailed, and corresponding use cases have been verified. Further, the open challenges and possible directions have also predicted and analyzed. We expect this study can promote the development of LEO multiple access possibilities in SAGIN.

References

[1] Z. Xiao, Z. Han, A. Nallanathan, O. A. Dobre, B. Clerckx, J. Choi, C. He, and W. Tong, “Antenna array enabled space/air/ground communications and networking for 6G,” IEEE J. Sel. Areas Commun., vol. 40, no. 10, pp. 2773–2804, Aug. 2022.
[2] Z. Jia, M. Sheng, J. Li, and Z. Han, “Towards data collection and transmission in 6G space-air-ground integrated networks: Cooperative HAP and LEO satellite schemes,” IEEE IoT J., vol. 9, no. 13, pp. 10 516–10 528, Oct. 2021.
[3] H. Cui, J. Zhang, Y. Geng, Z. Xiao, T. Sun, N. Zhang, J. Liu, Q. Wu, and X. Cao, “Space-air-ground integrated network (SAGIN) for 6G: Requirements, architecture and challenges,” China Commun., vol. 19, no. 2, pp. 90–108, Feb. 2022.
[4] L. He, J. Li, Y. Wang, J. Zheng, and L. He, “Balancing total energy consumption and mean makespan in data offloading for space-air-ground integrated networks,” IEEE Trans. Mob. Comput., pp. 1–14, Nov. 2022.
[5] Z. Xiao, J. Yang, T. Mao, C. Xu, R. Zhang, Z. Han, and X.-G. Xia, “LEO satellite access network (LEO-SAN) towards 6G: Challenges and approaches,” IEEE Wireless Commun., pp. 1–8, Dec. 2022.
[6] X. Cao, B. Yang, Y. Shen, C. Yuen, Y. Zhang, Z. Han, H. Vincent Poor, and L. Hanzo, “Edge-assisted multi-layer offloading optimization of LEO satellite-terrestrial integrated networks,” IEEE J. Sel. Areas Commun., pp. 1–1, Dec. 2022.
[7] Z. Jia, M. Sheng, J. Li, D. Zhou, and Z. Han, “Joint HAP access and LEO satellite backhaul in 6G: Matching game-based approaches,” IEEE J. Sel. Areas Commun., vol. 39, no. 4, pp. 1147–1159, Apr. 2021.
[8] C. Dong, Y. Shen, Y. Qu, K. Wang, J. Zheng, Q. Wu, and F. Wu, “UAVs as an intelligent service: Boosting edge intelligence for air-ground integrated networks,” IEEE Network, vol. 35, no. 4, pp. 167–175, Aug. 2021.
[9] M. S. Alam, G. K. Kurt, H. Yanikomeroglu, P. Zhu, and N. D. DÃ o, “High altitude platform station based super macro base station constellations,” IEEE Commun. Mag., vol. 59, no. 1, pp. 103–109, Feb. 2021.
[10] T. Ahmmed, A. Alidadi, Z. Zhang, A. U. Chaudhry, and H. Yanikomeroglu, “The digital divide in Canada and the role of LEO satellites in bridging the gap,” IEEE Commun. Mag., vol. 60, no. 6, pp. 24–30, Jun. 2022.
[11] Z. Jia, M. Sheng, J. Li, D. Zhou, and Z. Han, “VNF-based service provision in software defined LEO satellite networks,” IEEE Trans. Wireless Commun., vol. 20, no. 9, pp. 6139–6153, Sep. 2021.
[12] N. Cheng, F. Lyu, W. Quan, C. Zhou, H. He, W. Shi, and X. Shen, “Space/Aerial-Assisted Computing Offloading for IoT Applications: A Learning-Based Approach,” IEEE J. Sel. Areas Commun., vol. 37, no. 5, pp. 1117–1129, May 2019.
[13] B. Di, L. Song, Y. Li, and H. V. Poor, “Ultra-dense LEO: Integration of satellite access networks into 5G and beyond,” IEEE Wireless Commun., vol. 26, no. 2, pp. 62–69, Apr. 2019.
[14] S. Ji, M. Sheng, D. Zhou, W. Bai, Q. Cao, and J. Li, “Flexible and distributed mobility management for integrated terrestrial-satellite networks: Challenges, architectures, and approaches,” IEEE Network, vol. 35, no. 4, pp. 73–81, Jul. 2021.
[15] R. Wang, M. A. Kishk, and M.-S. Alouini, “Ultra-dense LEO satellite-based communication systems: A novel modeling technique,” IEEE Commun. Mag., vol. 60, no. 4, pp. 25–31, Apr. 2022.
[16] T. Ma, B. Qian, X. Qin, X. Liu, H. Zhou, and L. Zhao, “Satellite-terrestrial integrated 6G: An ultra-dense LEO networking management architecture,” IEEE Wireless Commun., pp. 1–8, Dec. 2022.
[17] X. Cao, B. Yang, C. Huang, C. Yuen, Y. Zhang, D. Niyato, and Z. Han, “Converged reconfigurable intelligent surface and mobile edge computing for space information networks,” IEEE Network, vol. 35, no. 4, pp. 42–48, Aug. 2021.
[18] P. CassarÃ¡, A. Gotta, M. Marchese, and F. Patrone, “Orbital edge offloading on mega-LEO satellite constellations for equal access to computing,” IEEE Commun. Mag., vol. 60, no. 4, pp. 32–36, Apr. 2022.
[19] Y. Zhang, C. Chen, L. Liu, D. Lan, H. Jiang, and S. Wan, “Aerial edge computing on orbit: A task offloading and allocation scheme,” IEEE Trans. Network Sci. Eng., pp. 1–11, Sep. 2022.
[20] H. Chen, M. Xiao, and Z. Pang, “Satellite-based computing networks with federated learning,” IEEE Wireless Commun., vol. 29, no. 1, pp. 78–84, Feb. 2022.
[21] L. He, B. Liang, J. Li, and M. Sheng, “Joint observation and transmission scheduling in agile satellite networks,” IEEE Trans. Mob. Comput., vol. 21, no. 12, pp. 4381–4396, Dec. 2022.
[22] Z. Xiao, T. Mao, Z. Han, and X.-G. Xia, “Near space communications: A new regime in space-air-ground integrated networks,” IEEE Wireless Commun., vol. 29, no. 6, pp. 38–45, Dec. 2022.
[23] Z. Jia, Q. Wu, C. Dong, C. Yuen, and Z. Han, “Hierarchical aerial computing for Internet of Things via cooperation of HAPs and UAVs,” IEEE IoT J., early access, Feb. 2022.
[24] Z. Kassas, J. Morton, F. van Diggelen, J. Spilker, and B. Parkinson, “Navigation from low earth orbit-part 2: Models implementation and performance,” in Position, Navigation, and Timing Technologies in the 21st Century. Wiley, 2021, vol. 2, pp. 1381–1412.
[25] R. M. Ferre, E. S. Lohan, H. Kuusniemi, J. Praks, S. Kaasalainen, C. Pinell, and M. Elsanhoury, “Is LEO-based positioning with mega- constellations the answer for future equal access localization?” IEEE Commun. Mag., vol. 60, no. 6, pp. 40–46, Jun. 2022.