Space-Air-Ground Integrated Network for Disaster Management: Systematic Literature Review

The occurrence of any kind of natural disaster will eventually lead to the loss of life and property. Countries where such disasters occur make every effort to monitor such disasters and aid as quickly as possible. However, in some cases, a rescue cannot be sent because no information is available to initiate any type of rescue operation. This is usually because common disaster management systems (DMS) use on board or ground networks to route information from the disaster scene to rescue headquarters (HQ)


Introduction
e mitigation of loss of human lives in any kind of disaster situation is one of the key aspects of any disaster management system.Whether a disaster occurs in the form of a tra c accident or in the form of a natural disaster, the major objective for rescuers is to lessen the loss of lives.To prevent lives in such situations, an e ective disaster management system (DMS) is signi cantly important.A DMS can be made highly e ective with the provision of high-quality imagery and data to conduct a successful rescue operation.
is data can be fetched through the use of technological applications and devices.As per [1], monitoring technologies such as sensor networks and unmanned aerial vehicles (UAV) can be employed for the reduction of disaster scenarios.
Disasters and hazards can occur at any time, whether they be natural or human-made and can result in the loss of lives.However, using information technology (IT) infused preventive techniques, researchers believe that such loss of lives can be entirely negated.Several recent research covers di erent aspects of IT applications in disaster management systems [2][3][4].Despite these advancements, there are still several challenges related to the assimilation of data pertaining to a disaster-struck area.One of the major issues in data accumulation during a disaster scenario is the timely communication of data to the DMS rescue team.Without such data communication the detection of disaster and prevention of loss is difficult.erefore, the use of air-based networks is a substantial research area for any type of disaster scenario.
Using UAV [5], it is possible to take aerial photos and then communicate this data to the disaster management (DM) headquarters for further processing.Additionally, vehicle networks can be used to obtain data from disasters [6][7][8].ough such systems are effective to a certain extent, however, the effectiveness can be increased with the use of satellites along with air and ground support.

Motivation.
e use of space, air, and ground networks has led to the specific field of the space-air-ground integrated network (SAGIN).SAGIN can attain more data relevant to the rescue operation of a disaster scenario effectively and efficiently.us, the utmost importance is to accumulate, review, and classify all of the research that has been performed in the area of SAGIN, especially in the case of disaster management.
Research studies for the advancement of disaster management using state-of-the-art techniques are available.However, the number of research papers related to SAGIN was extremely limited and the literature review of SAGIN for disaster management systems has not been conducted yet.Coupled with the use of such technology in the application of disaster managementa major problem is the lack of existing literature.
is is evident from Table1 where a majority of the high-scoring research is survey-based research, but there are only 4-5 survey research studies that have focused on making use of SAGIN in general for emergency situations or disaster relief.
e rationale behind this study is to motivate and help the researchers to know about SAGIN and its effectiveness in the disaster management system.Moreover, this study analyses the progress of SAGIN for disaster management systems.

Contribution.
is research provides a comprehensive systematic review of the complexities of SAGIN in the application of disaster management including different architectures and protocols.
e main contribution of this review is that it provides insight into existing disaster management systems that are making use of SAGIN.Moreover, this research work also highlights open research issues and challenges for any type of disaster scenario.Our results indicate that several challenges are faced by disaster management systems such as hardware-based challenges, network-based characteristics and communication protocols related challenges, availability and accuracy of imagery data, and security and privacy issues.eir solutions need to be further investigated for disaster management systems.

Background
e second type of disaster that can occur is humanmade, or more specifically any accident such as the collision of vehicles.e severity of this disaster in comparison to natural disasters is similar, however, the former can be avoided by following proper rules and regulations.erefore, there is also a need for an effective management system that can ensure reducing the risk of accidents efficiently.

Space-Air-Ground Integrated
Network.SAGIN was introduced to resolve the problems faced due to limited network capacity and coverage present in a ground communication system.It is expected that networks will provide more resources than the current situation to cope with the increasing demands in traffic and various other services.
e SAGIN architecture comprises three main segments: space, air, and ground.
ese segments can either work interoperationally or independently (as presented in Figure 1).In the space segment, the network is composed of constellations and satellites as well as their corresponding terrestrial infrastructures (ground stations).
ese constellations and satellites are present in different orbits and have different characteristics.
e air network segment can be considered an aerial mobile system that can use either aircraft for information transmission, procurement, and processing or UAVs or hot air balloons.And the ground network segment consists of terrestrial networks such as cellular networks, wireless local area networks (WLANS), or vehicular ad-hoc networks (VANETs).

Research Methodology
A systematic literature review (SLR) has been conducted for the purpose of examining the advancement in the disaster management area.e approach of SAGIN and ITS (Intelligent Transportation System) is closely examined and discussed in this research for their applicability in this area.
e main rationale for this study is to scrutinize and review different architectures and solutions for disaster management systems.To provide an effective SLR, this paper employs the methodology given in [138] for the provision of impartial research in the selection of data and the representation of its results.e research methodology is presented in Figure 2.
In the initial step of this SLR, the research objectives to determine the purpose of this survey were defined (as demonstrated in Figure 2).Afterward, these research objectives were transformed into the research questions.In the next step, all the existing research was gathered using a search string.After the collection process, screening was performed to filter out the research articles based on the keywords, abstract, and then the contents of the research article.In the final step, the data was extracted to answer the research questions.

Research Objectives.
is research provides attention to the following research objectives: e identification of research questions is required to provide the selection of research papers related to that specific question.e formulated research questions are presented in Table 2.

Search String.
e next step in this systematic literature review is the selection of different publication channels used to search for relevant articles pertaining to the defined research questions (RQs).After this, a search string is defined to search the previously defined publication channels.Using the defined search string, multiple research articles were collected, and then their results were aggregated to provide a solution to the RQs.
e selected publication channels have been specifically chosen that overly provide a close relationship between their scientific accuracy and the objects of this research.e selected publication database in which the search string has been applied ranges from IEEE to Science Direct.Complete detail of all the publication sources along with the search string are provided in Table 3.   Applied Computational Intelligence and Soft Computing

Screening for Relevant Results
. Not all the collected research articles were considered in this SLR as some were less related to the defined RQs.erefore, some results were filtered and excluded using the study provided by [139].
e provided study defines the screening of research articles while considering two phases.In the first phase, the collected research articles are filtered according to their titles.In the next phase, the filtered articles were further excluded/included based on their abstracts.e criteria which define the exclusion of research articles are as follows: (i) ose research articles that are not published in any journal, conference, or technical report.(ii) ere is no innovation or use of emerging ideas and technologies in the research articles.(iii) ose articles that have been published before 2010.
(iv) Research articles that had no relationship with the defined search string.
(v) Research papers that were not in the English language.
3.5.Quality Assessment.After the relevant research papers have been selected the next step in this SLR is to provide quality assessment (QA). is QA is conventionally performed to verify and evaluate the quality of the research articles collected and is performed using a questionnaire.However, for this SLR, the selected papers are assigned some quantitative values as per the mapping study presented by [140].e values that are to be assigned are provided below as well as in Table 4.To identify the different issues and challenges presented in the existing studies for SAGIN related systems

Analysis
In this section, a complete analysis of each research question, shown in Table 2, has been performed against different selected research articles.For all the selected primary research articles, their obtained QA score along with their publication name has been listed in Table 1.
A comprehensive analysis of a disaster management system that employs the use of SAGIN is a difficult challenge as it consists of different architectures and different other factors such as machine learning, artificial intelligence, and different types of network functions.Using the previously defined research questions, this research has made use of 50 research works from the last 10 years and was used to analyze the different research questions.Due to the limitation of fewer available works on the topic, we have short-listed research articles which were most relevant to find the research gap.

RQ1 Assessment: Key Publication Channels Targeted for DMS.
For the selection of different primary research articles, various publication channels were made use of.A list of the selected research papers along with their publication channel has been provided in Table 5.Also provided in the Table is the number of occurrences of research articles published in a particular journal (J) or conference (Conf.).

RQ2 Assessment: Frequency of Change in Research with
Respect to Time.
e quantity of the various selected research articles that have been published over the last decade is represented in Figure 3.It can be seen from the figure, that the majority of the research was conducted from 2018 to 2021 with the later years seeing more publications in this specific research area which ultimately points toward the growing interest in this research topic.

RQ3 Assessment: Approaches Used for Design of SAGIN
Based DMS.From the analysis of different research articles, it is found that different approaches in research have been made to design an effective DMS as shown in Figure 4. Inspired by the research work conducted in [141][142][143], the findings of some of the more recent works, have been summarized in Table 6.Each approach used in the selected research articles has been classified below:  109,112,115,116,118,120,125,126,129,132] for the concept of SAGIN overall or in the field of DMS.

Model.
Similarly, different authors have defined specific models for either the transmission characteristics or other network characteristics for the betterment of data transmission between the different components of SAGIN [19,30,50,51,57,60,71,78,80,87,97,110,114,119,124].4.3.6.Platform.Only 1 research article has focused on creating a platform for the monitoring of the different components or network characteristics of SAGIN [14,68,70].
Over the course of the last decade, different research has been conducted using the concept of SAGIN.However, not all the research has made use of the term SAGIN, as some have made use of the term space information network (SIN) or space transmission network (STN).e premise of all the terms is the same which is the use of geostationary satellites 6 Applied Computational Intelligence and Soft Computing     e evaluation demonstrated the effectiveness of the proposed scheme in terms of handover latency and average signalling cost [32] Use of space-arial-assisted end-edge-cloud orchestrated computing framework and the proposal of an alternating optimization algorithm To satisfy the computing of remote IoT applications by using a specialized cloud-based framework.Formulation of a fairnessaware resource scheduling problem to reduce the maximum task execution delay among IoT devices.Also, the proposal of an alternating optimizing.An algorithm based on the block coordinate descent method Neglection in the optimization of user computing delay in the worst scenario, which is the maximum computing delay of users

MATLAB R2012a
Evaluation through simulation where results showed an improvement in the reduction of computing delay of IoT devices.Also, an outperformance of the proposed end-edge-cloud orchestrated scheme in comparison to the endedge computing coordination method [42] Proposal of new network architecture with the combination of SAGIN and IoT To take advantage of SAGIN's large satellite coverage aside from the ground network propagation delay, a network architecture using SAGIN and IoT is proposed

None
Logical proposal of architecture [48] A framework for air traffic control based on SAGIN is proposed.satellite communication protocol specifications (SCPS) are proposed for air traffic communication Discussion on the issues faced by the aviation industry.Aside from this, a framework, ADS-B, incorporating the use of SAGIN is proposed which would ensure the safe operation of flights

Communication of airplanes and ATC centres
for the safe operation of flights STK Framework logically discussed.SCPS showed enhanced performance and proves itself viable to be used for space links Applied Computational Intelligence and Soft Computing and air and ground devices to create a network.e only difference that SIN and SAGIN have is that the former also makes use of sea networks.As such, this research has also included the research that uses these terms.Furthermore, this research has also devised a taxonomy that showcases the use of the different terms of SAGIN, SIN, and STN in relation to the mitigation of a disaster.is devised taxonomy is presented below in Figure 5. e taxonomy presents the different characteristics of disaster mitigation techniques as well as the different types of disasters and their locations where they normally occur.Some of the selected research work is focused on designing an architecture of SAGIN which can be then used for different types of scenarios and case studies.While some of the research works have made use of existing architecture and instead proposed solutions to a different aspect of SAGIN.
Murtaza and Jianwei [34] present those multiple architectures that have been proposed over the years incorporating the use of SAGIN, however, most of the proposed work is general abstracts or is based on a single dedicated application.
e authors present that for there to be an effective SAGIN architecture, the proposed work should be based on the TCP/IP stack.Another issue that the authors discussed was that due to the orbital positions of space nodes, the management of the dynamic network topologies is difficult.Hence, the authors have proposed a SAGINbased architecture that is based on the TCP/IP stack while also proposing a simple mechanism that would help in maintaining static IPs for all the space nodes (satellites).
Another research approach made was the integration of SAGIN with IoT conducted by [42] which not only provides large satellite coverage but also provides the different characteristics of ground network propagation delay.Another unique research approach conducted by [22] is the integration of artificial intelligence (AI) with SAGIN.e authors made this research in order to resolve the issues of data failures and computation slowdowns in SAGIN.As such, the authors proposed a novel coded storage-andcomputation architecture (CSC-AI) that provides reliable transmission of big data as well as fast computational offloads.
Some authors have focused their research on SIN and proposed different architectures.In [16], the authors have proposed an autonomous-based system (AS) architecture that has several components ranging from the use of GEO satellites as the backbone nodes of the network, LEO satellites for the purpose of enhanced coverage nodes, and the use of high-altitude platforms for the sake of meeting service requirements for emergency situations.e authors also proposed a hierarchal AS network model which basically divides the complex SIN network into smaller AS networks for easier control of the topology.Similarly, Yu et al. [58] proposed a system architecture for SIN as well as conducted a preliminary integrated demonstration environment for the proposed architecture.
Other authors have, however, focused their research on improving the architecture of SAGIN by incorporating either the use of vehicular networks [22,46,54] or software defined networks (SDN) [11,17,20,47] or mobile edge computing (MEC) or all these terms combined.
Some research has focused on the design or the model of different network protocols or characteristics that would ultimately help in the improvement of SAGINbased architecture.In [19], the authors have proposed an SDN-based SAGIN model as well as a new load-balancing dynamic routing algorithm for disaster scenarios that is based on the multidimensional resources and energy (LBMRE-OLSR), which through simulation showed a reduction in end-to-end delay and packet loss rate.Another model presented by [30] is based on the scenario of emergency logistics which is then used for a proposed novel scheme of cross-network radio scheduling in SAGIN based on a genetic algorithm and unified resources mapping.Other research articles that propose routing protocols include [9], which proposed a multipath routing algorithm based on network coding for resolving issues of high bit error rate and long propagation delays in SINs.
Another research area that has seen research on SAGIN is the security perspective.Different researchers have either focused on the encryption of data [26,28,43] or the trustworthiness of nodes [27,37] present in the network.

RQ4: Different Issues and Challenges to Overcome for an
Effective SAGIN Based DMS.After a comprehensive review of the selected papers, it can be presented that different solutions and characteristics have been proposed in relation to SAGIN and disaster management systems.However, though some of the papers have discussed the shortcomings and challenges presented in the research area, this paper has  Applied Computational Intelligence and Soft Computing identified some additional issues and challenges that need to be resolved for a proper SAGIN-based disaster management system.ese different challenges are also presented illustratively in Figure 6. e actual implementation of a SAGIN-based architecture can be difficult due to a critical scenario such as a disaster in a remote or a hilly area where even with a small landslide or snowfall the terrestrial infrastructure needed for a SAGIN can get damaged.

Networking Challenges.
Another issue that is also critical is the issue of network-based characteristics and routing/communication protocol.As in a disaster scenario, reliability as well as fast communication of the latest information related to that disaster is a key factor to be used in rescue operations.As such, an effective routing/communication protocol is needed for the effective dissemination of data.

Data Imagery.
As in any disaster situation, the first and foremost information required by a rescuer is the availability of images that show the current situation of the disaster area.However, the collection of multiple images as well as the transfer of images to the disaster Head Quarters (HQ) can prove to be difficult as the reliability of these images is very important as they would be used to launch a rescue operation.Any misinformation related to the imagery data can prove a disaster.

Security Challenges.
Another issue that poses a difficult challenge is the issue of security of either the entire architecture or structure of the SAGIN-based network or the encryption of data that is captured of the disaster scene and relayed to the HQ.In the situation of any such issue, the privacy of individuals caught in the disaster area would be at risk.From a network point of view, another security issue would be the misinformation provided due to trust issues of nodes being used in the SAGIN architecture.As such, an effective trust management scheme for all the nodes in either SAGIN or SIN is required.4.4.5.Software Issues.Security issues are not a current problem in the scenario of SAGIN-based DMS, however, this can be an evident issue in the future.Disaster mitigation in terms of SAGIN is an application used by end users for the reporting of any type of disaster.is would be a key aspect to get real-time information regarding a disaster occurrence in a certain location.Another aspect related to software issues is that with the introduction of AI in network architectures, it is also an important challenge to implement such a mechanism of cognition in SAGIN architecture to formulate a rescue operation as quickly as possible while keeping in consideration past disasters.

Quality Assessment Score.
e quality of selected articles is assured by the quality assessment criteria mentioned in the research methodology (subsection VII, Table 4), in the form of scores, as presented in Table 1.rough this assessment result, it would be possible for researchers to select their relevant papers according to their requirements.

Related Work
Disaster Management is a vast field of study and there are new approaches that are developed in different research studies.Many studies have been conducted to design new frameworks to improve this new paradigm of DM, but few survey studies and literature reviews have been performed to evaluate the research showing the new frameworks and approaches.ese approaches are reviewed in this section.e authors in [10] mainly presented the fundamental concept of SAGIN and 6G and key technologies of UAV and satellite-based communications.
ey also elaborated the 6G-SAGIN and its orientation towards the aspect of elemental design, as well as the concept of UAV-as-a-service to expand the comprehension of 6G-SAGIN.is research was presented in an elaborative and explanatory manner, however, there was little applicability discussed regarding the DMS.
e aerial platforms and the HSAT networks are surveyed in [29].ey discussed the technologies for disaster management.
ey outlined the challenges of the HSAT networks.However, the applications of the HSAT for DMS are not elaborated in the article.
SIN is observed to explicate the relationship between the services, security threats, and mechanisms in [41].Additionally, they presented the future direction of research in the security of SIN.ey explained the secure routing and In this research study, theuse of UAVs as Aerial Base Stations (ABS) is simulated for communication during such disasterous situation where the communication is uneven and broken [56].Effective communication probability (ECP) was used to evaluate the performance of a communication network, by measuring connectivity and throughput coverage.ey proposed better communication can be acquired through the flexible implementation of ABS.
e results show a noticeable increase in ECP when some ABS are deployed in optimal locations.is work is quite related to the scenario of the topic under consideration.However, the survey and comparison of the technique with other related studies can be extensive.
In [61], 6G-SAGIN from the aspect of a service-oriented network is reviewed to propose a service-oriented SAGIN management architecture.ey discussed cloud-edge synergy technologies and heterogeneous resource orchestration technologies.Additionally, the prospective directions are presented.A similar study on UAV communication with an aspect of 5G/B5G wireless networks is produced in [100], with a brief introduction of SAGIN, as well as related challenges of integrated networks are also comprehended.
is study entails an exhaustive review of several 5G techniques established on UAV platforms and the open issues and future directions for research.Still, the applications of SAGIN in DMS need some more substantial advancement.
To the best of our knowledge, there is no systematic literature review with a discussion of disaster management and SAGIN implementation techniques published yet with taxonomy and analysis.All the related research work on this area is summarized in Table 6, and the detailed comparison of other surveys with this study is summarized in Table 7.

Conclusions
e occurrence of disasters, either they be natural or humanmade can be damaging.erefore, there is an imperative need of a suitable and operationally effective disaster management system to alleviate the issues that arise from such disasters.Over the last decade, several investigations have been performed to overcome these concerns using UAV [20,32], sensor networks [144], and SAGIN [56].However, remote areas where ground networks are unable to reach can be responsible for more issues related to the provision of emergency response.In such situations, space networks which include the use of satellites and air networks, along with the use of aircraft and UAVs, integrated with a ground network (SAGIN) can solve the mentioned issues.is research provides a systematic literature review of 50 research articles, which constitute a comprehensive review of different research approaches adopted by the researchers.
e research gap we have found in this paper indicates that several challenges are faced by disaster management systems such as hardware-based challenges, network-based characteristics, communication protocols related challenges, availability and accuracy of imagery data, and security and privacy issues that need to be tackled and their solutions need to be further investigated.

AGIHN:
Air-ground integrated heterogeneous network ABS: Aerial base stations AI: Artificial intelligence AS: Autonomous based systems CA: Civil aircrafts CAA-SAGIVN: Civil aircrafts, augmented space-air-ground integrated vehicular networks CORE: Computing research and education CSC-AI: Coded storage-and-computation architecture DL: Deep learning DM: Disaster management DMS: Disaster management system ECP: Effective communication probability GEO: Geostationary earth orbit HetNet: Heterogeneous network HQ: Headquarters HSAT: Hybrid satellite-aerial-terrestrial IoT: Internet of ings IoV: Internet of vehicles IP: Internet protocol IT: Information technology ITS: Intelligent transportation system JCR: Journal

RO1:
To produce a thorough assessment of the latest research conducted in relation to disaster management systems RO2: To present different existing ITS-based disaster management systems for the case of human-made disasters RO3: To analyze the SAGIN-based architectures and solutions for disaster scenarios RO4: To identify different research gaps in terms of challenges and issues 3.2.Research Questions.

Figure 3 :Figure 4 :
Figure 3: Distribution of selected research papers by year.

Figure 6 :
Figure 6: Different challenges and issues.

4. 4 . 1 .
Hardware Challenges.Hardware-based challenges are one of the biggest as well as the most critical issue present.e reason for this deduction is that even with the introduction or proposal of different architectures or routing/communication protocols.

Table 1 :
Quality assessment score.

Table 3 :
Search string for the collection of papers from digital libraries.
"Disaster management" OR "emergency management" OR "rescue" OR "rescue operation" OR "accident management") AND ("space-air-ground network" OR "space-air-ground integrated network" OR "SAGIN" OR "space information network" OR "SIN" OR "STN") To evaluate this, the journal citation reports (JCR) and CORE ranking for computer science are considered for evaluation.e assigned score for either JCR or CORE is provided in Table4.e scoring of this evaluation is denoted by d.

Table 6 :
RQ3: algorithm/approaches used for an effective SAGIN based DMS.

Table 7 :
Comparison of this study to other studies.DL) based anomaly detection scheme.Overall, this paper summarizes the upcoming technologies and delivers inspiration for future exploration, though, the problems related to DMS were not extensively discussed.