A Comprehensive Survey on Orbital Edge Computing: Systems, Applications, and Algorithms
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Computer Science. The day in computer science: software, systems and theory on arXiv.
Tom: I'm Tom.
Jane: And I'm Jane, here with Lu, Meng and our in-house large language model, Lalam.
Tom: We're looking at the paper: "A Comprehensive Survey on Orbital Edge Computing".
Paper summary: Tom: So we're looking at this paper, "A Comprehensive Survey on Orbital Edge Computing: Systems, Applications, and Algorithms." What’s the main thing they’re telling us about orbital edge computing?
Jane: Well, basically, they are surveying the whole field. The core idea is that because we have so many satellites now and cheaper computer chips in them, we can push computing power up into space.
Lu: That's right. They define orbital edge computing as a network where the satellite itself acts as an edge node, which is different from just putting a small computer on the ground <ref:2306.00275#pg3>.
Meng: So what’s the big payoff here? What problem are they trying to solve with this whole concept?
Tom: The paper claims that this approach improves communication performance because you process sensor data right up in orbit, which leads to things like on-orbit data preprocessing and analyzing remote sensing images in real time <ref:2306.00275#pg4>.
Jane: It’s contrasting with mobile edge computing, which uses the movement of LEO satellites and limited resources to ease pressure on satellite communication <ref:2306.00275#pg5>. So, it's about using the space environment differently to save bandwidth back down to Earth.
Lu: The architecture relies on these satellite constellations, which offer better coverage and lower launch costs compared to older systems <ref:2306.00275#pg9>. They also point out that LEO satellites are particularly good because they have lower latency and cheaper launch costs than geostationary satellites <ref:2306.00275#pg10>.
Meng: Lower launch costs are important, but what about the actual hardware? What kind of stuff are these edge nodes running?
Tom: The system involves optimizing both the satellite hardware and its software architecture, focusing on how it communicates and computes <ref:2306.00275#pg2>. They mention using Commercial Off-the-Shelf devices with advanced chip manufacturing for implementing state-of-the-art AI algorithms while keeping power consumption low <ref:2306.00275#pg8>.
Jane: It sounds like the operating system on these spacecraft, things like Linux or SpaceOS, has to be really reliable because the space environment is so harsh <ref:2306.00275#pg9>. It's not just a simple cloud setup; it’s dealing with real physics and space constraints.
Lu: And that brings us into the algorithms they survey. They look at computation offloading, resource management, and mobility management as the three main areas of focus <ref:2306.00275#pg2>.
Paper summary: Tom: Computation offloading is modeled like a mixed integer nonlinear programming problem, and they use deep reinforcement learning methods like Deep Q-network to minimize task delay and energy consumption <ref:2306.00275#pg2>. That sounds computationally heavy.
Meng: If we were trying to deploy this, what’s the cost of running those kinds of AI algorithms on a satellite? What breaks at scale?
Lu: Resource management is often formulated as a mixed-integer programming problem, and they suggest using methods like the Lagrange multiplier method to find solutions <ref:2306.00275#pg2>. Mobility management then uses the predictable paths of satellites to assign tasks based on channel conditions <ref:2306.00275#pg2>.
Jane: So, for someone who just listens, this means that when you're using a service like ultra-long distance video conferencing or tracking an object, the processing isn't all happening down on Earth <ref:2306.00275#pg10>. It’s happening where the data is, which cuts down on latency.
Tom: Exactly. And they look at applications like object tracking and disaster surveillance, but also things like providing positioning and navigation services for the Internet of Vehicles in remote areas <ref:2306.00275#pg5>.
Meng: But the challenges must be huge if they're surveying this stuff. What are the real roadblocks they see?
Lu: They point out that resource-constrained satellite payloads and the complexity from large-scale satellite networks create NP-hard optimization problems <ref:2306.00275#pg2>. It’s hard to find a perfect solution when everything is constrained.
Jane: So, what are the future directions they suggest? Where is this going next?
Tom: They emphasize software-defined satellites, using inter-satellite laser links for better communication efficiency, and developing more accurate testbeds to simulate real space environments <ref:2306.00275#pg2>. The goal is autonomous task scheduling and efficient resource allocation in these complex scenarios <ref:2306.00275#pg2>.
Meng: Simulation is key here, right? How do you test something that’s in space without actually launching a satellite?
Jane: They stress that simulation tools are essential because you can't just plug it into the ground network simulator <ref:2306.00275#pg2>. They mention simulators like SNS3 and Hypatia, which model large low earth orbit constellations based on NS-three <ref:2306.00275#pg2>.
Paper summary: Lu: They also highlight testbeds like Celestial, which lets you run high-definition video conference tasks using Starlink satellites to get a realistic environment for evaluation <ref:2306.00275#pg2>. These tools help analyze how satellite configuration affects the algorithm performance <ref:2306.00275#pg2>.
Tom: So, we’ve covered the systems, the applications, and the algorithms they are looking at in this survey of Orbital Edge Computing. It's a broad look at what's possible <ref:2306.00275#pg1>.
Jane: And it really shows how feasibility is being established through designing these optimization algorithms and specialized testbeds <ref:2306.00275#pg2>. It’s moving from theory to something that can be tested in a more realistic setting.
Lu: The big implication for the field is showing that OEC has wide-ranging potential across space technology, agricultural remote sensing, the military, and even next-generation mobile communication <ref:2306.00275#pg2>.
Meng: For deployment right now, I’m still focused on those resource constraints and the complexity of scheduling tasks in these massive networks <ref:2306.00275#pg2>. We need to know what actually breaks when you try to scale this up <ref:2306.00275#pg1>.
Tom: So, the paper sets the stage by mapping out all the significant areas of focus, which is helpful for anyone trying to tackle these complex scheduling problems <ref:2306.00275#pg1>. It gives us a clear map of what’s been researched so far <ref:2306.00275#pg1>.
Jane: And ultimately, it suggests that OEC is expected to enable significant changes in the edge computing landscape and help grow the Internet of Things <ref:2306.00275#pg2>. It’s a big area for future development.
Lu: It really does suggest that satellite edge computing has proven its feasibility through the design of optimization algorithms and specialized testbeds <ref:2306.00275#pg2>. That’s a solid finding based on the survey <ref:2306.00275#pg1>.
Meng: We need to keep looking at those simulation tools and see if they can actually predict what happens when we push these models to real space conditions <ref:2306.00275#pg2>. That’s where the practical engineering work starts <ref:2306.00275#pg1>.
Tom: It seems like the survey itself is a really comprehensive resource for understanding the state of play in orbital edge computing right now <ref:2306.00275#pg1>. It’s a good starting point for anyone getting into this area <ref:2306.00275#pg1>.
Jane: That’s the gist of it. We’ve seen how the authors map out all these different technological components and challenges in this paper about Orbital Edge Computing <ref:2306.00275#pg1>.
Conclusion: Tom: So we're wrapping up this look at Orbital Edge Computing, which is really just a big overview paper by several authors <ref:2306.00275#pg1>.
Jane: It’s a comprehensive survey, Tom, meaning they’ve covered the whole landscape of these systems and applications.
Lu: Exactly. They went deep into the hardware, the software architecture, and all the math needed to make it work in orbit <ref:2306.00275#pg2>.
Meng: So what’s the main point they’re making at this conclusion about what this whole field means?
Tom: They're saying that orbital edge computing has proven its potential across a ton of different areas, from space tech to agricultural sensing and even mobile communication <ref:2306.00275#pg2>.
Jane: It seems they’re emphasizing that the feasibility of using satellites for edge computing isn't just theoretical anymore, it’s supported by real optimization algorithms and testbeds <ref:2306.00275#pg1>.
Lu: They point out that this approach is expected to make some significant changes to how we think about edge computing overall and how we grow the Internet of Things <ref:2306.00275#pg1>.
Tom: It’s a big shift from just thinking about edge computing on Earth to thinking about it up there in orbit <ref:2306.00275#pg1>.
Jane: And the authors suggest that we need to keep building those simulation tools and testbeds to make this actually happen in a real space environment <ref:2306.00275#pg1>.
Lu: Right, so they’ve mapped out the challenges—like those NP-hard optimization problems—and pointed toward the next steps for autonomous task scheduling <ref:2306.00275#pg2>.
Meng: So we’ve seen the scope of this work, but what does that mean for how we actually build these systems in the next few years?
Changhao Wu, Yuanchun Li, Mengwei Xu, Chongbin Guo, Zengshan Yin, Weiwei Gao, Chuanxiu Chi
Innovation Academy for Microsatellites of Chinese Academy of Science of China University of Chinese Academy of Sciences Beijing University of Posts and Telecommunications
cs.DC
Submitted: 2023-06-01
Updated: 2026-11-16
Comments: 18 pages, 9 figures and 5 tables
DOI: 10.1016/j.cja.2024.11.026
License: http://creativecommons.org/licenses/by-sa/4.0/
Importance score: 72/100
The gist: The paper provides a comprehensive survey of Orbital Edge Computing (OEC), covering its systems, applications…
Key concepts
- Orbital Edge Computing (OEC)
- OEC places computing resources directly onto satellites to act as edge nodes. This allows satellites to perform tasks like preprocessing sensor data in space, improving communication performance and enabling real-time remote sensing analysis.
- Contrast with Mobile Edge Computing (MEC)
- Unlike MEC, which focuses on mobile devices, OEC utilizes the mobility of LEO satellites and their limited resources. This approach is designed to alleviate pressure on satellite communication systems by processing data where it is collected.
- Computation Offloading
- This involves deciding whether to perform a computation on the satellite or send it to ground infrastructure. The paper models this as an optimization problem, often using deep reinforcement learning methods like DQN to minimize task delay and energy use.
- Simulation and Testbeds
- Because OEC involves unique space-ground interactions, simulation tools like SNS3 are used to model large satellite constellations. Platforms such as Celestial provide realistic testbeds for evaluating high-definition video tasks using actual satellites.
Terminology
Summary
The paper provides a comprehensive survey of Orbital Edge Computing (OEC), covering its systems, applications, and algorithms to address challenges arising from the explosion of satellites and COTS hardware enabling a new paradigm in computing <ref:2306. p. 0 1
A Comprehensive Survey on Orbital Edge
Computing: Systems, Applications, and Algorithms
Changhao Wu, Student Member, IEEE, Yuanchun Li, Member, IEEE,
Mengwei Xu, Member, IEEE, Chongbin Guo, Member, IEEE,
Zengshan Yin, Weiwei Gao, and Chuanxiu Chi
Abstract—The number of satellites, especially those operating
in low-earth orbit (LEO), is exploding in recent years. Additionally, the use of COTS hardware into those satellites enables
a new paradigm of computing: orbital edge computing (OEC).Where the paper says it
Definition and Context of OEC
Orbital Edge Computing (OEC) is defined as a network architecture where computing resources are pushed to satellites, allowing them to act as edge nodes in the network <ref:2306.00275#pg3>. This architecture improves communication performance by processing sensor data in orbit, leading to diversified services such as on-orbit data preprocessing and real-time remote sensing image analysis <ref:2306. p. 1 Due to the computation resources are located closer
to the data source, OEC greatly improves the communication
performance of the network by processing sensor data in orbit. OEC provide more diversified services, such as on-orbit data
preprocessing, real-time remote sensing image analysis and
disaster monitoring. This feature allows for vast design spaces
with multiple parameters, rendering several novel approaches
feasible. The mobility of LEO satellites in the network and limited
resources of communication, computation, and storage make it
challenging to design an appropriate scheduling algorithm for
specific tasks in comparison to traditional ground-based edge
computing.Where the paper says it
Where the paper says it
Network and Communication Aspects
The foundation of OEC architecture relies on satellite constellations, which offer superior coverage and lowered launch costs compared to traditional systems <ref:2306.00275#pg9>. LEO satellites are particularly attractive due to their better latency and lower launch cost, contrasting with the high access latency of GEO satellites <ref:2306. p. 1 The development of mobile communication motivates improvement in satellite communication with better performance. LEO satellites attract research attentions due to its better latency and lower launch cost. Satellite Communication: Frequency is a crucial factor
for satellite communication, microwave (300MHz-300GHz)
is suited to satellites because of its wide bandwidth and
high frequencies as well as the ability to use small antennas. Table II shows the frequency bands used by satellites. Inter-satellite communication performs better, especially longdistance, because light propagates faster in a vacuum than in
fiber [33]. Inter-satellite links (ISLs) include intra-orbit and
inter-orbit links [34].Where the paper says it
Where the paper says it
Where the paper says it
System Components and Hardware
The OEC system involves satellite hardware, software, and architecture optimization focusing on communication and computation <ref:2306.00275#pg2>. Satellite hardware often utilizes Commercial Off-the-Shelf (COTS) devices with advanced chip manufacturing processes to implement state-of-the-art AI algorithms with low power consumption <ref:2306. p. 6 The satellite needs to face
extreme difference in temperature and powerful radiation. The
commercial off-the-shelf (COTS) devices with advanced chip
manufacturing process such as 7nm are unable to cope with
single particle events in the space environment, which is
harmless to hardware but it can cause many logical errors
in software. Howerver, the Linux that applied in spacecraft is different from
that we used in personal computer (PC) or embedded device. SpaceX’s Linux is a self-made, deeply customized, distributed
operating system, which is not a third-party distribution.Where the paper says it
Where the paper says it
Key Applications in OEC
OEC has vast potential across several application domains, including object tracking and disaster surveillance <ref:2306.00275#pg5>. In AR/VR and UHD video applications, OEC provides ubiquitous access with ultra-long distance and low-latency transmission to overcome terrestrial network limitations <ref:2306. p. 8 As an effective solution, OEC provides ubiquitous access and ultra-long distance
with low-latency transmission, opening up further possibilities
for these applications [79]. Celestial [18] evaluated the performance of WebRTC video
meetings using OEC based on the Starlink network. Their proposed
fine-grained joint offloading and caching scheme leverages the
orbit-ground collaboration to provide communication services
for vehicles in rural areas, isolated islands, and remote places
where terrestrial infrastructure is not available. The scheme
aims to provide real-time Enhanced Cellular Satellite Assisted
Global Navigation System (EC-SAGINs) services for terrestrial vehicles. The data generated is growing rapidly in
terms of size and variety, which creates immense pressure on
space-ground communication [82]. However, OEC can analyze
this data, provide near real-time analysis, and significantly enhance decision-making processes.Where the paper says it
Where the paper says it
Where the paper says it
Algorithms for Optimization
Several critical optimization problems in OEC require advanced algorithms, including computation offloading, resource management, and mobility management <ref:2306.00275#pg2>. Computation offloading is often modeled as a mixed integer nonlinear programming problem, solved using deep reinforcement learning-based methods like Deep Q-network (DQN) to minimize task delay and energy consumption <ref:2306.00275#pg2>. Resource management can be formulated as a mixed-integer programming (MIP) problem, with solutions often achieved using the Lagrange multiplier method <ref:2306.00275#pg2>. Mobility management leverages the predictable trajectories of satellites to assign tasks based on channel conditions <ref:2306. p. 0 The mobility of LEO satellites in the network and limited
resources of communication, computation, and storage make it
challenging to design an appropriate scheduling algorithm for
specific tasks in comparison to traditional ground-based edge
computing. This article comprehensively surveys the significant
areas of focus in orbital edge computing, which include protocol
optimization, mobility management, and resource allocation.Where the paper says it
Challenges and Future Directions
Significant challenges remain, including resource-constrained satellite payloads and the complexity arising from large-scale satellite networks which present NP-hard optimization problems <ref:2306.00275#pg2>. Future research directions emphasize software-defined satellites, inter-satellite laser links for enhanced communication efficiency, and the development of more accurate testbeds to simulate real space environments <ref:2306. p. 2 3
Paper
outline
Ⅱ Orbital Edge Computing
Ⅳ Applications of OEC
Ⅴ Simulators and Testbeds
Ⅵ Algorithms
Issues, Challenges and Ⅶ Future Research Directions
Satellite Constellations, Networking and Communication
Mobile Computing: From Ground to Space
Concept of OEC
Advanced Technology in Space
Orbital Edge Computing VS Terrestrial Edge Computing
Satellite Federated Learning
Earth Observation
AR/VR and UHD video
Internet of Vechicles and Autonomous Driving
Scientific Research
Satellite Constellations for Simulation
Experiment Methods
Simulation Tools
Computing Offloading
Resource Managemt
Mobility Managemt
Issues and Challenges
Opportunities and Future Research
Ⅲ Systems
Hardware
Software
Architecture
Fig. The outline of this paper. As satellite tasks become more complex, this method
is inefficient. In the future, autonomous task scheduling could
alleviate these limitations, enabling satellites to make decisions
online with minimal manual intervention.Where the paper says it
Where the paper says it
Simulation and Testbed Development
Simulation tools are essential for evaluating OEC performance due to the distinction between satellite and ground networks <ref:2306.00275#pg2>. Popular simulators include SNS3 and Hypatia, which are designed to simulate large-scale low earth orbit constellations based on NS-3 <ref:2306. p. 0 LEO satellite network is a promising technology due to
its ubiquity and invulnerability [7]. To address this issue,
large-scale low earth orbit (LEO) satellite network has been
extensively investigated [8]. Celestial [18] evaluated the performance of WebRTC video
meetings using OEC based on the Starlink network. Three
users sent a high-definition video stream with 2.Where the paper says it
Where the paper says it
Conclusion
OEC has demonstrated wide-ranging potential in diverse fields, including space technology, agricultural remote sensing, the military, the Internet of Things, and next-generation mobile communication <ref:2306.00275#pg2>. The feasibility and effectiveness of satellite edge computing have been proven through the design of optimization algorithms and specialized testbeds <ref:2306. p. 2 3
Paper
outline
Ⅱ Orbital Edge Computing
Ⅳ Applications of OEC
Ⅴ Simulators and Testbeds
Ⅵ Algorithms
Issues, Challenges and Ⅶ Future Research Directions
Satellite Constellations, Networking and Communication
Mobile Computing: From Ground to Space
Concept of OEC
Advanced Technology in Space
Orbital Edge Computing VS Terrestrial Edge Computing
Satellite Federated Learning
Earth Observation
AR/VR and UHD video
Internet of Vechicles and Autonomous Driving
Scientific Research
Satellite Constellations for Simulation
Experiment Methods
Simulation Tools
Computing Offloading
Resource Managemt
Mobility Managemt
Issues and Challenges
Opportunities and Future Research
Ⅲ Systems
Hardware
Software
Architecture
Fig. The outline of this paper.Where the paper says it
References
[1] A. Fox, R. Griffith, A. Joseph, R. Katz, A. Konwinski, G. Lee, D. Patterson, et al., “Above the clouds: A berkeley view of cloud computing,” Dept of Electrical Eng and Comput Sciences UCB/EECS 2009; <ref:2306. p. 7 to the LEO constellation, and the edge computing LEO (LEC)
controller has a global view for task scheduling. Furthermore,
the LEC controller can determine whether the task should
be offloaded to the cloud [13].Where the paper says it
[2] Y. C. Hu, M. Patel, D. Sabella, N. Sprecher, and V. Young, “Mobile edge computing—a key technology towards 5g,” ETSI white paper vol 11 no 11 pp 1–16;
[3] S. Naveen and M. R. Kounte, “Key technologies and challenges in iot edge computing,” in 2019 Third international conference on I-SMAC (IoT in social, mobile, analytics and cloud)(I-SMAC) pp 61–65;
[4] B. Cao, J. T. Zhang, X. Liu, Z. H. Sun, W. X. Cao, R. M Nowak, and Z H Lv, “Edge-cloud resource scheduling in space-air-groundintegrated networks for internet of vehicles” IEEE Internet of Things Journal vol 9 no 8 pp 5765–5772;
[5] F. Yang, S. Wang, J. Li, Z. Liu, and Q Sun, “An overview of internet of vehicles” China communications vol 11 no 10 pp 1–15;
[6] A. Al-Ansi, A M Al-Ansi, A Muthanna, I A Elgendy, and A Koucheryavy, “Survey on intelligence edge computing in 6g: Characteristics, challenges, potential use cases and market drivers” Future Internet vol 13 no 5 p 118;
[7] Q. Tang, Z Fei, B Li, and Z Han, “Computation offloading in leo satellite networks with hybrid cloud and edge computing” IEEE Internet of Things Journal vol 8 no 11 pp 9164–9176;
[8] J Liu, Y Shi, Z M Fadlullah, and N Kato, “Space-air-ground integrated network: A survey” IEEE Communications Surveys & Tutorials vol 20 no 4 pp 2714–2741; <ref:2306.00275#pg9>
[9] Y Henri, “The oneweb satellite system” Handbook of Small Satellites Technology Design Manufacture Applications Economics and Regulation pp 1–10;
[10] V L Foreman, A Siddiqi, and O De Weck, “Large satellite constellation orbital debris impacts: Case studies of oneweb and spacex proposals” in AIAA SPACE and Astronautics Forum and Exposition p 5200;
[11] L Wood, Y Lou, and O Olusola, “Revisiting elliptical satellite orbits to enhance the o3b constellation” arXiv preprint arXiv:1407.2521;
[12] T Kim and J P Choi, “Performance analysis of satellite server mobile edge computing architecture” in 2020 IEEE 92nd Vehicular Technology Conference (VTC2020-Fall) pp 1–6;
[13] C Li, Y Zhang, R Xie, X Hao, and T Huang, “Integrating edge computing into low earth orbit satellite networks: Architecture and prototype” IEEE Access vol 9 pp 39126–39137;
[14] M J Holzinger and M K Jah, “Challenges and potential in space domain awareness” Journal of Guidance, Control, and Dynamics vol 41 no 1 pp 15–18;
[15] S Wang, Q Li, M Xu, X Ma, A Zhou, and Q Sun, “Tiansuan constellation: An open research platform” in 2021 IEEE International Conference on Edge Computing (EDGE) pp 94–101;
[16] K Kanev and N Mirenkov, “Satellite cloud computing” in 2011 IEEE Workshops of International Conference on Advanced Information Networking and Applications pp 147–152;
[17] B Denby and B Lucia, “Orbital edge computing: Machine inference in space” IEEE Computer Architecture Letters vol 18 pp 59–62;
[18] T Pfandzelter and D Bermbach, “Celestial: Virtual software system testbeds for the leo edge” in 23rd ACM/IFIP International Middleware Conference (Middleware ’22) 2022;
[19] J Wei, J Han, and S Cao, “Satellite iot edge intelligent computing: A research on architecture” Electronics 2019;
[20] F Wang, D Jiang, S Qi, C Qiao, and H H Song, “Fine-grained resource management for edge computing satellite networks” 2019 IEEE Global Communications Conference (GLOBECOM) pp 1–6;
[21] J Fang, H Liu, J Sun, T Li, Y Liu, Z Guo, S Zhuang, and L Geng, “Learning-based task offloading in dynamic orbital edge computing network” in 2021 IEEE 23rd Int Conf on High Performance Computing & Communications; 7th Int Conf on Data Science & Systems; 19th Int Conf on Smart City; 7th Int Conf on Dependability in Sensor, Cloud & Big Data Systems & Application (HPCC/DSS/SmartCity/DependSys) pp 495–502;
[22] S Liu, X Hu, and W Wang, “Deep reinforcement learning based dynamic channel allocation algorithm in multibeam satellite systems” IEEE Access vol 6 pp 15733–15742;
[23] M Luo and J Feng, Satellite Communications.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI system can perform object tracking by utilizing a framework where
the LEC controller has a global view for task scheduling
to determine whetherthe task should be offloaded to the cloud
or anedge computing LEO (LEC) controller.
-
Improved AI system can enhance disaster surveillance by employing a framework where satellites in the same orbit plane are referred to as
formation flying,
allowing them to establish stable communication links and offload tasks between satellites. -
Improved AI system can optimize remote sensing data transmission by implementing an architecture where
OEC organizes satellite constellations to minimize communication consumption
and Bradley et al.'s strategy ofidentifying the desired signals and only transmitting them.
-
Improved AI system can improve AR/VR quality-of-service by leveraging OEC, as experiments showed that
satellite servers have the potential to significantly improve the quality-of-service (QoS) for clients in latency-sensitive applications
with end-to-end latencies below 16 ms. -
Improved AI system can enhance autonomous driving capabilities by using satellite edge computing as an auxiliary solution to provide a
more stable and reliable AI decision-making scheme for autonomous driving
by providinglower latency and greater computing resources.
-
Improved AI system can optimize resource allocation through game theory by addressing the problem of resource management in OEC, which is currently noted as having
relevant no work attempting to use game theory to address the resource management problem in OEC.
-
Improved AI system can reduce communication costs and guarantee data privacy by incorporating
the recognition of signal modulation in OEC and satellite FL
into its algorithms. -
Improved AI system can enable efficient data analysis in environmental monitoring by applying satellite federated learning, which allows for the analysis of
multiple sources, such as satellite imagery, ground sensors, and weather stations.
Abstract
The number of satellites, especially those operating in low-earth orbit (LEO), is exploding in recent years. Additionally, the use of COTS hardware into those satellites enables a new paradigm of computing: orbital edge computing (OEC). OEC entails more technically advanced steps compared to single-satellite computing. This feature allows for vast design spaces with multiple parameters, rendering several novel approaches feasible. The mobility of LEO satellites in the network and limited resources of communication, computation, and storage make it challenging to design an appropriate scheduling algorithm for specific tasks in comparison to traditional ground-based edge computing. This article comprehensively surveys the significant areas of focus in orbital edge computing, which include protocol optimization, mobility management, and resource allocation. This article provides the first comprehensive survey of OEC. Previous survey papers have only concentrated on ground-based edge computing or the integration of space and ground technologies. This article presents a review of recent research from 2000 to 2023 on orbital edge computing that covers network design, computation offloading, resource allocation, performance analysis, and optimization. Moreover, having discussed several related works, both technological challenges and future directions are highlighted in the field.
Sources
- Revisiting elliptical satellite orbits to enhance the O3b constellation
- AsyncFLEO: Asynchronous Federated Learning for LEO Satellite Constellations with High-Altitude Platforms
- FedSpace: An Efficient Federated Learning Framework at Satellites and Ground Stations
- Federated Learning in Satellite Constellations
- FedHAP: Fast Federated Learning for LEO Constellations Using Collaborative HAPs
Related papers
- Programmable Probabilistic Computer with 1,000,000 p-bits
- BatchGen: An Architecture for Scalable and Efficient Batch Inference
- Seamless Reconfiguration for DAG-BFT (Technical Report)
- ELDR: Expert-Locality-Aware Decode Routing for PD-Disaggregated MoE Serving
- iScheduler: Reinforcement Learning-Driven Continual Optimization for Large-Scale Resource Investment Problems
- TIRx: A Unified Tile Primitive ML Compiler Across Heterogeneous Backends