Satellite communication (SatCom) systems operations centers currently require high human intervention, which leads to increased operational expenditure (OPEX) and implicit latency in human action that causes degradation in the quality of service (QoS). Consequently, new SatCom systems leverage artificial intelligence and machine learning (AI/ML) to provide higher levels of autonomy and control. Onboard processing for advanced AI/ML algorithms, especially deep learning algorithms, requires an improvement of several magnitudes in computing power compared to what is available with legacy, radiation-tolerant, space-grade processors in space vehicles today. The next generation of onboard AI/ML space processors will likely include a diverse landscape of heterogeneous systems. This manuscript identifies the key requirements for onboard AI/ML processing, defines a reference architecture, evaluates different use case scenarios, and assesses the hardware landscape for current and next-generation space AI processors.

This paper provides a comprehensive survey on the application and development of Artificial Intelligence (AI) and Machine Learning (ML) in satellite communication (SATCOM). It explores the increasing integration of AI/ML technologies in SATCOM systems, highlighting their potential to enhance performance, efficiency, and adaptability in response to growing demands for connectivity and data processing. The survey categorizes various use cases across different layers of satellite networks, detailing conventional solutions and the advantages of employing AI/ML techniques. It discusses the challenges associated with onboard processing, including hardware constraints, radiation tolerance, and the need for efficient resource management. Furthermore, the document examines the role of neuromorphic computing and COTS (Commercial Off-The-Shelf) devices in facilitating AI applications in space environments. Finally, we discuss the long-term developments of AI in the SATCOM sector and potential research directions. Overall, the survey emphasizes the transformative impact of AI/ML on the future of SATCOM, paving the way for innovative solutions in next-generation satellite networks.

Very high throughput satellite (VHTS) systems are expected to have a huge increase in traffic demand in the near future. Nevertheless, this increase will not be uniform over the entire service area due to the non-uniform distribution of users and changes in traffic demand during the day. This problem is addressed by using flexible payload architectures, which allow the allocation of payload resources flexibly to meet the traffic demand of each beam, leading to dynamic resource management (DRM) approaches. However, DRM adds significant complexity to VHTS systems, so in this paper we discuss the use of one reinforcement learning (RL) algorithm and two deep reinforcement learning (DRL) algorithms to manage the resources available in flexible payload architectures for DRM. These algorithms are Q-Learning (QL), Deep Q-Learning (DQL) and Double Deep Q-Learning (DDQL) which are compared based on their performance, complexity and added latency. On the other hand, this work demonstrates the superiority a cooperative multiagent (CMA) decentralized distribution has over a single agent (SA).

Onboard satellite communication systems generate and manage coverage beams over the Earth. Depending on data traffic requirements, the number of beams, side lobe levels, nulls, and EIRP, their beamwidth must be efficiently generated and managed. Therefore, this paper describes an approach for beam pattern synthesis applied to geostationary satellite communication systems. The beam pattern synthesis can generate beams with a beamwidth variation from 0.45° to 1.5°, which can be controlled independently for the two principal cuts. In addition, other requirements have been considered, e.g., latitude, and longitude, required EIRP, minimum and maximum side love levels for the two principal cuts, and nulling direction. The output of the synthesizer is a weight matrix with beamforming coefficients of the required beam. The direct radiating array in this contribution utilizes an open-ended waveguide antenna as unit cell elements with a period of 0.875λ ₀ designed to work in left-hand circular polarization in the frequency band from 17.7 to 20.1 GHz. Since this design is intended for high-data rates applications, the minimum beamwidth requirements are very narrow. Therefore, 36 × 36 sub-arrays of 4 × 4 unit cells with a period of 3.5λ are considered to accomplish the beamwidth requirements while maintaining reduced computational and time resources for the weight matrix calculation compared to the conventional counterpart of 144×144 unit cells. The results show that the algorithm, which uses the surrogate optimizer, can compute the weight matrix and synthesize the beam with a slight deviation from the input data.

Maritime transport has become important due to its ability to internationally unite all continents. In turn, during the last two years, we have observed that the increase of consumer goods has resulted in global shipping deadlocks. In addition, the future goes through the role of ports and efficiency in maritime transport to decarbonize its impact on the environment. In order to improve the economy and people’s lives, in this work, we propose to enhance services offered in maritime logistics. To do this, a communications system is designed on the deck of ships to transmit data through a constellation of satellites using interconnected smart devices based on IoT. Among the services, we highlight the monitoring and tracking of refrigerated containers, the transmission of geolocation data from Global Positioning System (GPS), and security through the Automatic Identification System (AIS). This information will be used for a fleet of ships to make better decisions and help guarantee the status of the cargo and maritime safety on the routes. The system design, network dimensioning, and a communications protocol for decision-making will be presented.

The latest Satellite Communication (SatCom) missions are characterized by a fully reconfigurable on-board software-defined payload, capable of adapting radio resources to the temporal and spatial variations of the system traffic. As pure optimization-based solutions have shown to be computationally tedious and to lack flexibility, Machine Learning (ML)-based methods have emerged as promising alternatives. We investigate the application of energy-efficient brain-inspired ML models for on-board radio resource management. Apart from software simulation, we report extensive experimental results leveraging the recently released Intel Loihi 2 chip. To benchmark the performance of the proposed model, we implement conventional Convolutional Neural Networks (CNN) on a Xilinx Versal VCK5000, and provide a detailed comparison of accuracy, precision, recall, and energy efficiency for different traffic demands. Most notably, for relevant workloads, Spiking Neural Networks (SNNs) implemented on Loihi 2 yield higher accuracy, while reducing power consumption by more than $100\times $ as compared to the CNN-based reference platform. Our findings point to the significant potential of neuromorphic computing and SNNs in supporting on-board SatCom operations, paving the way for enhanced efficiency and sustainability in future SatCom systems.