Space-Air-Ground Integrated Network with Native Intelligence (NI-SAGIN): Concept, Architecture,Access and Networking

Ubiquitous coverage is one of the most important goals for mobile communication networks.To achieve this, integration of space, air, and ground networks is highly demanded, which expects to become the one of the enabling technologies for 6G networks.However, the integrated network faces the demand of orders of magnitude improvement in coverage, access capability and network service capability, which could not be achieved under the existing access and management mechanisms.To support this, artificial intelligence (AI) can be applied to provide technical means for solving the problems of organization and management of the space-air-ground integrated network.By virtualizing and “softwarizing” the resources of the entire network, and provisioning intelligence in the core network elements of the integrated network, the space-air-ground integrated network will be able to form native intelligence.The native intelligence will facilitate a highly autonomous network, which has the capabilities of self-adaptive scale, self-allocation resource, self-learning behaviors and self-evolution function, and is expected to meet diversified, refined and differentiated service quality requirements in the future.

The design of a native intelligent space-air-ground integrated network faces many challenges including but not limited to: space-air integrated network architecture, the evolution of network native intelligence, deployment of intelligent network elements in the control plane of the heterogeneous dynamic network, native intelligence based access control/mobile management/connection management, unified novel air interface, end-to-end virtualization technology, service-driven flexible slice customization, etc.The rapid development of various technologies, such as AI and SDN/NFV, big data, digital twins, cloud computing/edge computing/fog computing, has provided theoretical and technical support for solving the above problems.It has been proven that the eye-catching AI technologies, encompassing unsupervised learning, reinforcement learning, transfer learning, and federal learning, are able to enhance the overall performance of communication network systems.Furthermore, under the dual drive of "big data + digital twins", the corresponding and related design of AI architecture/algorithms/processes and information network architecture/algorithms/processes can constitute the basic technical system of the space-air-ground integrated network for future 6G.

The primary goal of this feature topic is to present the state-of-the-art original research and latest advances and innovations in for space-air-ground integrated network (SAGIN), including architecture,evolution, deployment and implementation of network native intelligence, and AI-driven integrated network operation management, e.g., access control,mobile management, connection management, slicing management, etc.All of the submitted papers are evaluated according to the standard reviewing process of China Communications.Following a rigorous peer-review process, 15 articles are accepted in this special issue.

The accepted papers cover topics about the architecture, access, networking, data processing, routing strategies and positioning in the satellite network,aerial network and Space-Air-Ground Integrated Network.We hope this special issue may inspire the existing and important future research works in related fields.

The space-air-ground integrated network (SAGIN)is regarded as the key approach to realize global coverage in future network and it reaches broad access for various services.Being the new paradigm of service, immersive media (IM) has attracted users’attention for its virtualization, but it poses challenges to network performance, e.g.bandwidth, rate, latency.However, the SAGIN has limitations in sup-porting IM services, such as 4K/8K video, virtual reality,and interactive games.In the article entitled “Service Customized Space-Air-Ground Integrated Network for Immersive Media: Architecture, Key Technologies, and Prospects,” a novel service customized SAGIN architecture for IM applications (SAG-IM)is proposed, which achieves content interactive and real-time communication among terminal users.State-of-the-art research is investigated in detail to facilitate the combination of SAGIN and service customized technology, which provides end-to-end differentiated services for users.Besides, the functional components of SAG-IM contain the infrastructure layer, perception layer, intelligence layer, and application layer, reaching the capabilities of intelligent management of the network.Moreover, to provide IM content with ultra-high-definition and high frame rate for the optimal user experience, the promising key technologies on intelligence routing and delivery are discussed.The performance evaluation shows the superiority of SAG-IM in supporting IM service.Finally, the prospects in practical application are high-lighted.

The article by Wang et al., “Holistic Service-based Architecture for Space-Air-Ground Integrated Network for 5G-Advanced and Beyond,” proposes a novel Holistic Service-based Architecture (H-SBA)for SAGIN of 5G-Advanced and beyond, i.e., 6G.The H-SBA introduces the concept of end-to-end service-based architecture design.The “Network Function Service”, introduced in 5G SBA, is extended from Control Plane to User Plane, from core network to access network.Based on H-SBA, the new generation of protocol design is proposed, which proposes to use IETF QUIC and SRv6 to substitute 5G HTTP/2.0 and GTP-U.Testing results show that new protocols can achieve low delay and high throughput, making they are promising candidate for H-SBA.

In the article entitled “Civil Aircraft Assisted Space-Air-Ground Integrated Networks: Architecture Design and Coverage Analysis,” a novel AI-enabled space-air-ground integrated networks architecture (SAGIN) is proposed.The proposed architecture consisted of LEO satellites and civil aircrafts carrying aerial base stations, called “civil aircraft assisted SAGIN(CCA-SAGIN)”.The assistance of civil aircrafts can reduce the stress of satellite networks, improve the performance of SAGIN, decrease the construction cost and save space resources.Taking the Chinese mainland as an example, the distribution of civil aircraft is analyzed, and characteristics of civil aircraft assisted networks (CAAN) is obtained.Taking Starlink as the benchmark, the service gap of CAAN is calculated, and the joint coverage constellation is designed.The simulation results prove that the number of satellites in CAA SAGIN can be greatly reduced with the assistance of civil aircrafts at the same data rate.

In the article entitled “ESMD-Flow: An Intelligent Flow Forwarding Scheme with Endogenous Security based on Mimic Defense in Space-Air-Ground Integrated Network,” the authors propose an intelligent flow routing scheme based on Mimic Defense(ESMD-Flow) to guarantee the security requirement of flow.With the help of software defined network(SDN), the scheme is aware of reliability of each node and link and adopt multipath routing strategy to ensure flows to be sent along the most reliable multiple paths.Furthermore, a multi-protocol forwarding strategy for realizing the multi-protocol dynamic forwarding of flows is developed based on the programming data plane.The test results demonstrate that the proposed scheme can improve the average path reliability for routing and increase the difficulty of network eavesdropping while improving the network throughput and reducing the average packet delay.

In the article entitled “Intelligent Passive Detection of Aerial Target in Space-Air-Ground Integrated Networks,” proposes an intelligent passive detection method for aerial target with the help of SAGIN, i.e.,satellites, and in turn offers a useful procedure of detecting aerial targets in the intricate SAGIN environment.Specifically, delayed feedback networks are utilized to refine the direct signals from the satellite in the reference channels.In addition, the satellite direct wave interference in the monitoring channels adopts adaptive interference suppression using the minimum mean square error filter.Furthermore, decoupling echo state networks are utilized to predict the clutter interference in the monitoring channels and construct the detection statistics accordingly.Finally, a multilayer perceptron is adopted to detect the echo signal after interference suppression.The simulation results demonstrate the effectiveness of the proposed method.

The rapid development and continuous updating of the mega satellite constellation (MSC) have brought new visions for the future 6G coverage extension,where the global seamless signal coverage can realize ubiquitous services for user terminals.However,global traffic demands present non- uniform characteristics.Therefore, how to ensure the on-demand service coverage for the specific traffic demand, i.e.,the ratio of traffic density to service requirement per unit area, is the core issue of 6G wireless coverage extension exploiting the MSC.In the article entitled“6G Service Coverage with Mega Satellite Constellations,” the authors discuss the open challenges to reveal the future direction of 6G wireless coverage extension from the perspective of key factors affecting service coverage performance, i.e., the network access capacity, space segment capacity and their matching-relationship.Then, they elaborate on the key factors affecting effective matchings of the aforementioned aspects, thereby improving service coverage capability.

Resource allocation is an important problem influencing the service quality of multi-beam satellite communications.In multi-beam satellite communications, the available frequency bandwidth is limited, user requirements vary rapidly, high service quality and joint allocation of multi-dimensional resources such as time and frequency are required.It is a difficult problem needs to be researched urgently for multi-beam satellite communications, how to obtain a higher comprehensive utilization rate of multi-dimensional resources, maximize the number of users and system throughput, and meet the demand of rapid allocation adapting dynamic changed the number of users under the condition of limited resources, with using an efficient and fast resource allocation algorithm.In the article entitled“Multi-Objective Deep Reinforcement Learning Based Time-Frequency Resource Allocation for Multi-Beam Satellite Communications,” the authors establishes a multi-objective optimization model based on the maximum the number of users and system throughput joint optimization goal, and proposes a multi-objective deep reinforcement learning based time-frequency two-dimensional resource allocation (MODRL-TF) algorithm to adapt dynamic changed the number of users and the timeliness requirements.Simulation results show that their algorithm could provide higher comprehensive utilization rate of multi-dimensional resources, and could achieve multi-objective joint optimization, and could obtain better timeliness than traditional heuristic algorithms, such as genetic algorithm (GA) and ant colony optimization algorithm (ACO).

In the article entitled “Fast and Accurate LEO Satellite-Based Direct Position Determination Assisted by TDOA Measurements,” a time difference of arrival assisted DPD (TA-DPD) which minimizes the searching area by the time difference of arrival measurements and their variances is proposed.In this way, the size of the searching area is determined by both geometrical constraints and qualities of received signals, and signals with higher SNRs can be positioned more efficiently as their searching areas are generally smaller.Both two-dimensional and three-dimensional passive localization simulations using the proposed TA-DPD are provided to demonstrate its efficiency and validity.The superior accuracy performance of the proposed method, especially at low SNRs conditions, is also verified through the comparison to conventional two-step methods.Providing a larger margin in link budget for satellite-based vessel location acquisition, the TA-DPD can be a competitive candidate for trusted marine location service.

In the article entitled “Joint Task Scheduling, Resource Allocation, and UAV Trajectory under Clustering for FANETs,” a new layered flying ad-hoc networks (FANETs) system of mobile edge computing (MEC) supported by multiple UAVs is established, where the first layer of user UAVs can perform tasks such as area coverage, and the second layer of MEC UAVs are deployed as flying MEC sever for user UAVs with computing-intensive tasks.In this system, the user UAVs are firstly divided into multiple clusters, and the tasks of the cluster members (CMs) within a cluster are transmitted to its cluster head (CH).Then, each CH’s tasks are executed locally execution (i.e., task scheduling).An optimization problem with the aim of minimizing the overall energy consumption of all user UAVs is formulated, under the constraints of task completion deadline and computing resource, which is a mixed integer non-convex problem and hard to solve.Next, an iterative algorithm is proposed by applying block coordinate descent methods.To be specific,the task scheduling between CH UAVs and MEC UAVs, computing resource allocation, and MEC UAV trajectory are alternately optimized in each iteration.For the joint task scheduling and computing resource allocation subproblem and MEC UAV trajectory subproblem, branch and bound method and continuous convex approximation technique are employed to solve them, respectively.Extensive simulation results validate the superiority of their approach to several benchmarks.

In the article entitled “Joint Scheduling and Resource Allocation for Federated Learning in SWIPT-Enabled Micro UAV Swarm Networks,” the authors develop a joint client scheduling and wireless resource allocation of federated learning (FL) in a simultaneous wireless information and power transfer (SWIPT) enabled micro UAV swarm network, where the BS transfers both ML model and power simultaneously to the UAVs, and the UAVs depend on the harvested and battery-stored energy to perform ML model training and uploading.For the purpose of improving the learning performance, a UAV scheduling and wireless resource allocation problem is formulated and two algorithms are proposed to solve the problem, i.e., the optimal and sub-optimal algorithm.The numerical results show that the sub-optimal algorithm outperforms the existing baselines and cam reach a near-optimal result.

In the article entitled “Joint Power-Trajectory-Scheduling Optimization in a Mobile UAV-enabled Network via Alternating Iteration,” they aim to minimize the energy consumption of the UAV-enabled mobile edge computing (MEC) system, where the UAV exhibits caching, computing and relaying capabilities to periodically provide services to cellular users and D2D receiver nodes.Through jointly optimizing the scheduling strategies of the user,transmitting power of the UAV and D2D-TX nodes,and UAV trajectory, the energy consumption and throughput can be guaranteed at the same time.Further, an alternating iteration algorithm, together with the successive convex approximation, penalty function, and Dinkelbach method are employed to transform the original problem into a group of solvable subproblems and the convergence of the method is proved.Simulation results show that the proposed scheme performs better than other benchmark algorithms, particularly in terms of balancing the tradeoff between minimizing UAV energy consumption and maximizing throughput.

Smart containers have been extensively applied in the maritime industry with the development of Internet of Things (IoT).But the limited coverage of the onshore base station (BS) can influence data offloading rate and delay in the offshore region.In this regard, “UAV-Assisted Data Offloading for Smart Container in Offshore Maritime Communications,”they propose a UAV-assisted data offloading model where the mobility of container vessel in the offshore region is considered.Further, based on the model a data offloading algorithm is proposed to reduce the average offloading delay under data-size requirements and available energy constraints of smart containers.Specifically, the convex-concave procedure is used to update time-slot assignment,offloading approach selection, and power allocation in an iterative manner.Simulation results show that the proposed algorithm can efficiently reduce average offloading delay and increase offloading success ratio.

The article entitled “Predictive Decision and Reliable Accessing for UAV Communication in Space-Air-Ground Integrated Networks,” they develop a predictive decision algorithm for reliable UAV communication in SAGIN with handover strategy to ensure the seamless connectivity.Firstly, the authors propose the concept of air controlling center to collect and update global information of UAVs, and to make decisions on selecting the optimal access points for UAV accessing.Then the optimization model that maximizes the overall link capacities with the limitation of the maximal average delay threshold of these links is considered.Based on the aforementioned model, the authors propose predictive decision algorithm to decide on the sets of optimal access points both at current and predictive future time as the output results.Using the results,they have then proposed the handover strategy for UAVs to maintain the approximately always-on connections with SAGIN infrastructures by switching handover states seamlessly.Simulation results have demonstrated that the proposal outperforms the existing methods and achieves reliable UAV communication against high mobility, sparse distribution, and physical obstacles.

In the article entitled “Routing Protocol for Heterogeneous FANETs with Mobility Prediction,” a routing protocol has been proposed for UAVs which utilizes UAVs’ mobility information in heterogenous FANETs.First, an SDN-based FANET architecture is given and based on it, an Extended Kalman Filter(EKF) for accurate mobility estimation and prediction of UAVs is applied.With the aid of mobility information, the routing problem is formulated as a graph decision problem and a Directional Particle Swarming Optimization (DPSO) approach is utilized to solve the problem.The extensive simulation results demonstrate that the proposed routing scheme can exhibit superior performance in improving the goodput, packet delivery ratio, and delay.

Last but not the least, it is a trend that utilizing UAVs in safe autonomous navigation systems to explore unknown environments where UAVs are equipped with multiple sensors to perceive the surroundings.However, data dissemination is a hard issue for UAV application in safe navigation.To tackle the problem, “Safe Navigation for UAV-enabled Data Dissemination by Deep Reinforcement Learning in Unknown Environments,” a new approach based on the most advanced dueling double deep Q network (dueling DDQN) with multi-step learning is developed, for the purpose of minimizing the whole weighted sum of the UAV’ s task completion time while satisfying the data transmission task requirement and the UAV’ s feasible flight region constraints.Specifically, the extra labels are added to the primitive states to improve the learning performance.The numerical results indicate the validity and performance superiority of the proposed algorithm under different data thresholds compared with two other benchmarks.

In conclusion, the Guest Editors of this feature topic would like to thank all the authors for their valuable contributions and the anonymous reviewers for their constructive comments and suggestions.We also would like to acknowledge the guidance from the editorial team of China Communications.