Spiral Blue’s Space Computer Ready for Launch
A new prototype Earth Observation computer, developed by Australia’s Spiral Blue, is set to be launched on a Falcon 9 rocket in December 2020, onboard the SW1FT satellite.
Australian-based start-up, Spiral Blue, have signed an agreement to launch their first prototype, space-based computer into orbit at the end of this year with Polish space manufacturer SatRevolution.
The prototype computer, called Space Edge Zero (SEZ) will be hosted on SatRevolution’s SW1FT satellite, currently scheduled for launch aboard a SpaceX Falcon 9 rocket for December 2020. The mission will be used to run a full end-to-end test of the onboard computer during 2021, demonstrating its technical capabilities, algorithms, and functionality whilst orbiting Earth.
“SpaceX is the top launch company globally and we’re excited to be flying alongside their Starlink satellites, whilst progressing well on our side,” said Taofiq Huq - founder, and CEO of Spiral Blue.
“We’re also excited that we're adding in the ability to update software after launch. If we do this right, we'll be able to add functionality after launch,” he said.
“In the long run, this will help keep aging satellites up to date with the latest algorithms and processing methods. For this launch specifically though, this means we'll be able to send up code to run new projects for customers or for research purposes even after launch,”
The Earth Observation Game
Earth Observation (EO) is a game-changer for humanity. When satellites first went into orbit and turned their cameras not towards the stars, but back down to Earth – the images revolutionised everything.
Philosophically we marvelled at the ultimate birds-eye beauty of our marble planet, teeming with life below. Commercially we were able to build applications and platforms that millions of people use on a daily basis. Strategically, we saw an opportunity to monitor, learn, and model our changing environment.
The Australian Space Agency, along with other government and science bodies has recognised the need for accurate, high-resolution, real-time EO data to be one of the pillars of building Australia’s space capabilities – especially given the opportunity to utilise this data for agricultural and water management and more drastically for air quality, fire, and natural disaster monitoring.
Images, or data, captured by EO satellites are transmitted back to Earth to be processed, using radio waves from the satellite down to terrestrial receiving antennas and stations. Whilst this has been working well for decades, the increased demand for real-time space-based data across different services is placing pressure on two fronts.
Firstly, the transmissions are limited to certain bands of radio wavelengths – as this part of the electromagnetic spectrum is shared by a large population, such as radio astronomers, submarine transmissions, wi-fi networks, military, and civilian ground and air operations, radar, etc. With a finite number of bands available for satellite communications and increasing demand for service - having a highly efficient transmission process in place is ideal, ensuring that as much data is being sent from the satellite to station in the quickest amount of time.
There are other portions of the spectrum being utilised for transmission purposes, that are currently being researched, such as the usage of lasers and x-rays to encode data before sending it from space back down to Earth - but this has not yet been sustainably achieved nor applied to commercial services.
The second pressure is the data itself. The need for higher resolution in images/data is becoming more useful – esp. when coupled with Global Positioning System (GPS) services – for example, self-driving cars need the most up-to-date and highest resolution images to provide the greatest degree of accuracy when performing their functions.
However, the cost of higher resolution data comes at a price of larger files – which in turn take longer to transmit over limited bandwidth between satellite and receiver stations. Thus, this makes the transmission process heavy and inefficient.
Space Edge Zero
Here’s where Spiral Blue steps in. The Sydney-based company is developing the Space Edge Zero, a computer that resides on satellites that takes images from optical instrumentation (like high-resolution cameras) and processes these images on board the satellite, prior to relaying them back down to Earth.
Using a mixture of classical remote sensing methods and modern artificial intelligence (AI) techniques, raw image data is converted into information aboard the satellite, reducing the file size by a range of 20 – 1,000 times, whilst keeping all the required information from the image/data file.
This provides a resolution to the second problem described above in terms of overcoming issues with bandwidth, in addition to making costs more efficient, imagery lead-time shorter, and overall improving the productivity of the process.
Orbiting onboard the SW1FT satellite launching later this year, SatRevolution will have a camera that is able to take EO images at 6 metre resolution. As part of the testing of the SEZ, these images will be passed to the onboard computer for processing in real-time, before they are relayed back down to Earth.
This opportunity allows the Spiral Blue team to test their process end-to-end: image capture + onboard processing + efficient Earth transmission.
Taofiq and his team are first and foremost looking at ensuring the SEZ gets to orbit and commences operations - a major milestone for any organisation.
“The most important thing for us is to see if it turns on. If it does, we'll have space heritage for our company, and for the SEZ. From there, we'll be looking to understand how far we can push SEZ's performance in orbit,” he said.
“We'll have initial results within a month after launch. From there, we should have a slow trickle of new learnings as we run the SEZ more and more over time, with new algorithms and applications.”
Spiral Blue was founded in 2017 and is made up of a three-person team based in Sydney – space engineer, Taofiq Huq (CEO), AI scientist Henry Zhong (Head of AI), and system/electrical engineer James Buttenshaw (Head of Hardware).
“I gained an interest in the concept of space startup after interning with Saber Astronautics a few years ago and then seeing the incredible success of SpaceX. From there, I wanted to join the space industry but wasn't keen on travelling overseas. I ultimately decided I would help build the industry in Australia. I knew I wanted to do earth observation, as it intuitively made sense to me as a necessary technology to help tackle climate change and other global problems. Spiral Blue was born from my trying to find new and interesting ways to approach earth observation,” said Taofiq.
The SEZ became a reality in August 2019, when Spiral Blue was awarded the Minimal Viable Product Grant by the New South Wales Government. Since then, the company has been putting the SEZ through its paces, including simulating the conditions the computer will experience in space.
“We plan to apply our learnings from Space Edge Zero to develop Space Edge One. This will take a couple of years at least, so until then we will be making small tweaks to SEZ for any missions we do until Space Edge One is ready. We're starting to evaluate future launch opportunities to test SEZ further, and to potentially use it to start delivering earth observation services to customers.”
To do so, the computer was run through thermal testing in a vacuum and placed into the Cobalt-60 Gamma Irradiator at the Australian Nuclear Science and Testing Organisation (ANSTO), ensuring that the radiation in space would not fry the onboard components.
The team also took out the Meteor Award for the most innovative use of space capability at the 2019 inaugural Gravity Challenge, hosted by Deloitte and Amazon Web Services. Partnering with Spectral Aerospace, the team proposed a new product called ‘Line of Insight’ – an application that provides daily forecasts of fire risks near-critical network infrastructure by meshing together real-time Internet of Things sensory data, satellite imagery, and weather data.