On a mission to create the world’s most secure and fastest global point-to-point low-latency orbital network, Rivada Space Network has been employing digital twins to create a simulation environment to build the same.
“Digital twin is fantastic,” said Declan Ganley, the founder and CEO of Rivada Space, in an exclusive interaction with AIM at the recently concluded AWS Summit in Washington DC.
“When you are working with your enterprise and government customers, you can show them all sorts of options like geo-fencing and routing. You can mesh every single vessel,” he added.
The Vital Digital Twin
Rivada’s digital twin enables them to accurately model the behaviour of their satellite constellation. This modelling capability is crucial for testing various operational scenarios before actual deployment.
By simulating different conditions, such as orbital dynamics and network traffic patterns, Rivada can predict how their satellite network will perform in real-world scenarios. This pre-testing reduces the risk of errors and enhances the efficiency of network deployment and operation.
“What we can do is, in AWS, build a digital twin of this entire constellation to be able to not only visualise the constellation and calculate visibility windows and contacts in seconds. This would usually take hours, but we get it down to seconds.
“That enables Rivada to run what-if scenarios, capacity planning, and after-support and using the serverless power of the cloud,” said Alistair McLean, principal tech BDM for aerospace and satellite, AWS.
Speaking about the tech configuration for the digital twin, McLean explains that they use a queue-based system, Amazon’s SQS (simple queue service) which creates Lambdas and Batches. “If you flip any good simulator on its head and actually feed it data, it really becomes a digital twin,” said McLean.
Elaborating on the vitality of digital twins in space and satellite technologies, Ganley talks about its usage on the manufacturing side. “You can do things in one day that five years ago would have taken you 50,” he said. Simply put, satellite efforts that require 100 years are easily cut to two years.
With this accelerated pace, Rivada Space, a relatively new entrant setup in 2022, is expected to bring low-earth-orbit satellites in 2028.
Interestingly, Ganley believes cost is a better deal than the time advantage. “Digital twins give you savings that are better than that,” he said.
Rivada’s approach is not the first time a satellite network company has adopted the concept of digital twin. Elon Musk’s satellite internet constellation Starlink has already adopted the same.
Space and Simulation
The use of digital twin technology in this sector not only helps create virtual models of satellites but also enables better security measures that facilitates smoother operations. They even help with troubleshooting and quicker resolution for network issues and disruptions.
The process thereby helps save time, cost and resources.
SpaceX has been using digital twins to test their Starlink and is said to be so advanced that it can even emulate the effects of solar flares on its network.
It was said that digital twins can reduce the time needed to deploy AI-driven capabilities by up to 60%, cutting capital and operating expenditures by up to 15%, and improving commercial efficiency by 10%.
Further, digital twins facilitate virtual testing within a secure and regulated environment, leading to substantial reductions in the time and resources needed for validation and verification.
Not Challenge Free
Though promising, digital twin also come with its challenges. It can encounter limitations such as integrating complex data sources, ensuring data accuracy and availability, scaling for large systems, addressing security concerns, and achieving interoperability in diverse environments.
However, a recent research paper named ‘Plotinus: A Satellite Internet Digital Twin System’, talks about a digital twin based on microservices and designed for satellite internet emulation.
Plotinus aims to tackle SAGIN (space-air-ground networks) development challenges with modular design for aerial vehicle emulation, flexible path computation methods, real-time emulation with live network traffic, and dynamic satellite network modelling.
Big Tech Leads the Way
GPU giant NVIDIA has been advancing its digital twin capabilities through various platforms and collaborations. Its Omniverse platform has been widely adopted across robotics and autonomous industries. Considering how simulated training is one of the key methods for training robots, digital twin is the perfect fit.
“Eventually we’ll have sophisticated autonomous robots working alongside humans in settings like kitchens — manipulating knives and other dangerous tools. We need digital twins of the worlds they are going to be operating in, so we can teach them safely in the virtual world before transferring their intelligence into the real world,” said Rev Lebaredian, the VP of omniverse and simulation technology at NVIDIA.
NVIDIA has also partnered with Foxconn to build AI factories or factory digital twins. “Our digital twin will guide us to new levels of automation and industrial efficiency, saving time, cost and energy,” said Foxconn chairman Young Liu.
Similarly, NVIDIA’s Earth is a digital twin cloud platform for simulating and visualising weather and climate conditions. The company has a similar project to monitor global environmental conditions via digital twin for the National Oceanic and Atmospheric Administration (NOAA).
Other industries are following suit. Semiconductor companies have already seen the need to employ digital twins to replicate their semicon factory settings. While this is in motion, it is possible that digital twins can help India catch up on the semiconductor race against incumbent players.