Aside from bolstering access to the final frontier, IBM also announced an open-source project to develop more advanced space debris tracking systems leveraging machine learning.
Access to orbit around Earth was once limited to a handful of space agencies around the globe. With the proliferation of spacefaring technologies and cost-efficient craft, low-Earth orbit (LEO)—the sliver of space extending to 1,200 miles above our planet—is now an increasingly populous mix of private and public interests. Today, LEO is brimming with government craft, commercial programs, university undertakings, venture capital funding, and more.
On Thursday, IBM announced two open-source projects in an effort to “democratize access” to space technologies and help track the debris field orbiting overhead. We spoke with Naeem Altaf, IBM Distinguished Engineer and CTO of space tech, to learn more about these programs.
Building more advanced space debris tracking systems
Since the launch of Sputnik more than 60 years ago, orbit around Earth has accumulated a large debris field filled with millions of objects ranging from untrackably small leftovers from past launches to school bus-sized objects. These objects whirl overhead at approximately 18,000 mph. During our conversation, Altaf referenced this expansive debris field as well as the risk of a Kessler Syndrome event; a potentially catastrophic situation in which one collision sets off a cascading event of subsequent collisions in orbit.
SEE: TechRepublic Premium editorial calendar: IT policies, checklists, toolkits, and research for download (TechRepublic Premium)
One of IBM’s latest open-source announcements is focused on designing more advanced space debris tracking systems. Currently, different governments around the globe use tracking systems to monitor space debris in orbit, however, the exact locations of different objects will vary depending on which entity one speaks with, according to Altaf.
“If you ask them for the data, they all come up with a really different picture… and that is a huge concern,” Altaf said.
Atlaf used the flight control tower analogy and likened the scenario to a flight tracker that suddenly begins to show a flight with two or three different paths. As one would imagine, this would present myriad safety and logistical challenges. Additionally, as IBM points out, it’s not completely understood how other factors such as space weather affect the trajectories of space debris.
To assist, IBM has created the Space Situational Awareness (SSA) project, operated in partnership with the University of Texas at Austin, which leverages two models to monitor space debris. The physics-based SSA model incorporates Cowell’s formulation to model “perturbation” space debris orbit “caused by the Earth.”
A second model uses machine learning to predict errors in orbit predictions using XGBoost gradient-boosted regression trees, per the IBM release. With USSTRATCOM data acting as “the ground truth,” the machine learning model is trained using the physical model’s orbit predictions to predict errors in the physics model.
Satellite swarms and “CubeSats as a service”
The second IBM open-source project is designed to enable greater access to “satellite swarms.” These swarms are in essence a number of satellites operating in tandem as part of a larger collective objective. In recent years, more organizations are beginning to use smaller CubeSats as opposed to much larger traditional satellites. Altaf likens the shift to evolutions in terrestrial software development.
“We used to build these big, large monolithic software. Then came the concept of microservices. So instead of having this big monolithic thing, I can divide it into 10 different functions and I can have those small programs, which are independent, loosely coupled, and they can do their things, right? Instead of me spending so much money [and] effort in building this one thing, I can start building in chunks and they can do a different function for me,” Altaf said.
IBM’s Kubesat open-source project creates a cognitive, autonomous satellite framework, enabling “simulation and optimization of multi-satellite communications,” according to the company. Satellite swarm communication and collective cooperation can function autonomously enabling a swarm to collaborate or separate depending on the objective, per IBM.
“Basically you are forming a logical cluster on the fly in that swarm. And then it goes and performs the function and once that event is finished, they all go back to their state, independent state, where they can be available for any other job or task to be given. So, [it’s] sort of like a distributed computing and batch processing jobs are happening,” Altaf said.
Depending on the sensors onboard and capabilities, swarms could be called upon for a host of applications and users ranging from meteorological analysis to deforestation, according to Altaf, likening the concept to “CubeSats as a service.”
“It becomes a multitenant system where different jobs are being submitted by different entities, so it can basically cater to different needs,” Altaf said.
SEE: 5 Internet of Things (IoT) innovations (free PDF) (TechRepublic)
IBM has open-sourced the KubeSat Cognitive Autonomous Framework to enable a larger number of operators to utilize these technologies. Additionally, IBM announced that it would also open source the code developed in partnership with Stanford University undergraduates. These extensions simulate “satellite-to-satellite, satellite-to-ground station, and satellite to ground sensor communications over a meshed NATS messaging platform,” per IBM.
View original article here Source