Feature: Space electronics
T e reason for the long transmission times lies in the source of
the data: colour images obtained in space oſt en have a resolution of 25-130 megapixels in video mode, and a single uncompressed image easily generates 75MBytes of data. At 30 frames per second, this would require a transmission bandwidth of 18Gbit/s. T e bandwidth of a connection to a nanosatellite orbiting
the Earth in 400-600km can reach Gbit/s but is generally only around 50Mbit/s, so demand by far exceeds performance. T e larger satellites can generate terabytes of data per day; satellite communication ground stations are spread around the Earth, each prioritised diff erently, which adds latency to the data chain.
Solving space-data latencies By installing a space cloud directly in the satellite, Swedish company Unibap is addressing the problem of data latency. Our solution, called SpaceCloud, uses artifi cial intelligence (AI) to decide locally which data is relevant. Instead of sending massive raw data streams, it only transmits necessary data to the ground station, such as analysis results and positioning commands; see Figure 1. “Unibap’s SpaceCloud delivers a fl exible yet powerful
infrastructure for artifi cial intelligence in space,” said Dr. Fredrik Bruhn, former guest researcher at NASA, founder and former CEO of Unibap. “T e platform integrates cloud services, intelligent data processing, sensor management, data storage and data analytics, plus on-demand transmission of relevant information.” T ere are several reference apps that demonstrate how cloud
Cloud computing in space By Dr. Fredrik Bruhn, Director, Unibap
G
eostationary satellites are 36,000 kilometres away in space, yet our everyday systems for navigation, communication, weather forecasts, traffi c monitoring, television, and a lot more all rely on them. In traffi c-monitoring applications, satellite
sensors capture image data to be sent to and evaluated by Earth stations. Most of the time, this approach works smoothly, as response times are suffi ciently long and not every second counts. However, when evaluating data in near real time, say in cases of potential air traffi c collision, the limits of what is possible are quickly reached.
32 March 2022
www.electronicsworld.co.uk
Figure 1: Unibap’s SpaceCloud can evaluate data from images depicting road or air traffi c in near real time, which makes it possible to react instantly to critical situations
technology can be used in space, for example to do advanced image and video compression, analyse scientifi c data, perform precision agriculture and recognise vehicles, ships or aircraſt using machine learning. T e development environment uses pre- analysed satellite images for this. In principle, the development system is also suitable for any deep-learning application. Unibap’s SpaceCloud made its maiden fl ight aboard the
D-Orbit’s ION WILD RIDE mission on 30 June 2021; see Figures 2 and 3.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46
