In A First-Of-Its-Kind Space Experiment, Amazon Web Capabilities (AWS) Was Able To Successfully Execute Its AWS Compute And Machine Learning Services On An Orbiting Satellite

Amazon Web Services (AWS) made the announcement today at AWS re:Invent 2022 that it successfully launched a package of AWS computing and machine learning (ML) software on an orbiting satellite, in what is believed to be the first experiment of its sort in space. The experiment was carried out in low Earth orbit (LEO) over the course of the past ten months. Its purpose was to test a way that would enable customers to collect and analyse critical space data directly on their orbiting satellites utilising the cloud in a quicker and more effective manner.
The first time that AWS edge capabilities have been provided onboard an orbiting satellite, customers have been given the ability to automatically analyse massive volumes of raw satellite data while the satellite is in orbit and only downlink the most useful images for storage and further analysis. This has resulted in significant cost savings and improved customers’ ability to make timely decisions.
“Using AWS software to perform real-time data analysis onboard an orbiting satellite, and delivering that analysis directly to decision makers via the cloud, is a definite shift in existing approaches to space data management. [This] analysis can be delivered directly to decision makers in real time. According to Max Peterson, vice president of the worldwide public sector for Amazon Web Services (AWS), “It also helps push the frontiers of what we feel is possible for satellite operations.” “By providing strong and secure cloud capability in orbit, we enable satellite operators the opportunity to interface with their spacecraft in a more effective manner and deliver updated directives via AWS tools with which they are already accustomed.”
Amazon Web Services (AWS) is dedicated to overcoming the technical obstacles that come with operating in space, such as networks with limited bandwidth and high latency. AWS worked together with two of its global space partners, D-Orbit and Unibap, to directly solve these difficulties as they relate to satellite operations.
A pioneer in the field of space logistics and transportation services, D-Orbit is also a participant in the Amazon Web Services (AWS) Partner Network (APN). D-Orbit was able to quickly analyse vast amounts of space data directly onboard its orbiting ION satellite by utilising the computation and machine learning resources offered by Amazon Web Services (AWS) and applying them to Earth Observation (EO) imagery.
AWS Partner Network (APN) Partner D-Orbit is a market leader in the space logistics and transportation service industry. D-Orbit is also an APN Partner. D-Orbit was able to quickly analyse massive amounts of space data directly onboard its orbiting ION satellite by utilising the computation and machine learning resources offered by Amazon Web Services (AWS) and applying them to Earth observation (EO) imagery.
According to Sergio Mucciarelli, vice president of commercial sales at D-Orbit, “Our customers want to securely process increasingly enormous amounts of satellite data with very low latency.” [Citation needed] [Citation needed] “This is something that is restricted by making use of older technologies, such as downlinking all of the data in order to process it on the ground.” We believe in the push toward edge computing, and we believe that it can only be done with space-based infrastructure that is suitable for the task at hand. This will give customers a high degree of confidence that they will be able to run their workloads and operations reliably in the harsh environment of space operation.
The teams worked together to develop a software prototype that would incorporate the tools that they had collectively determined to be necessary for the EO mission. These tools include AWS ML models to analyse satellite imagery in real time and AWS IoT Greengrass to provide cloud management and analytics even during periods of limited connectivity. The prototype was built through a collaborative effort between the teams.
Unibap, a high-tech firm based in Sweden that is also an AWS Partner, was responsible for the construction of the space-qualified processing payload that the AWS software prototype was put onto. After that, the processing payload for Unibap was mounted onto a D-Orbit ION satellite and sent into orbit around the earth. The crew achieved its first successful contact with the payload on January 21, 2022, and also carried out the first remote command sent from Earth to space at that time. A couple of weeks later, the group started carrying out its trials.
Dr. Fredrik Bruhn, Chief Evangelist in Digital Transformation at Unibap and Co-Founder, stated that the company’s goal is to “help customers quickly turn raw satellite data into actionable information that can be used to disseminate alerts in seconds, enable onboard federated learning for autonomous information acquisition, and increase the value of data that is downlinked.” “We want to help customers quickly turn raw satellite data into actionable information that can be used to disseminate alerts Users will be able to acquire more timely insights and optimise how they use their satellite and ground resources if they are given real-time access to AWS edge services and capabilities on orbit, which will be provided by Amazon Web Services.
Throughout the course of the experiment, the group used a variety of ML models to analyse the sensor data from the satellite in order to rapidly and automatically identify particular objects both in the sky, such as clouds and smoke from wildfires, and on the ground, such as buildings and ships.
Because raw satellite photos and datasets of this kind are typically rather huge, the team devised a method to divide the massive data files into several more manageable pieces. When using the AI and ML services offered by Amazon Web Services (AWS), image file sizes can be reduced by up to 42 percent, resulting in increased processing speeds and the ability to make real-time inferences in orbit. In order to accommodate for a greater delay tolerance between communications, the team was responsible for managing the bidirectional transit of space data via several ground station links. In order to accomplish this goal, a dependable TCP/IP proxy had to be managed between the satellite and the AWS Cloud. The ground workers were able to more easily manage the automatic file transfers thanks to this improvement, which eliminated the need for them to manually process the downlinks across several connections.
Today, the collaborative experiment is still being conducted in space, where AWS, Unibap, and D-Orbit are continuing to test new capabilities that go beyond the initial set of test objectives. For instance, the group would like to investigate additional techniques for processing raw data while the satellite is in orbit, as well as more refined methods for the distribution of data. AWS and its partners are continuing their investigation into ways to close the technological gaps that exist in LEO, and the data and insights gleaned from these ongoing on-orbit experiments are essential to this effort.
According to Peterson, “Ultimately, AWS believes that by giving customers the ability to evaluate their space data quickly and securely using AWS on-orbit will help them make important decisions more quickly.”
Discover more about the ways in which AWS for Aerospace and Satellite helps its customers and partners all over the world.