St John’s in SpaceKate
Over the last 9 months, the St John’s Year 9 Computing Club has been taking part in the European Astro Pi Mission Space Lab Challenge. Students were tasked with designing and coding an experiment to run on the International Space Station which would investigate one of two broad areas: Life in Space, where they investigate life inside the Columbus module using sensors onboard a Raspberry Pi computer and display some form of output using an LED matrix, or Life on Earth where they investigate life on the Earth’s surface using either an infrared camera or a visible-light camera deployed in front of an Earth-facing window on the ISS.
St John’s had four teams that completed Phase 1 of the challenge and submitted a design to the judges at the end of October. Three of those teams made it through to Phase 2 where they then got to code and test their experiments, before deploying their code on the ISS in April, as part of Phase 3. Phase 4 is now complete, the teams have received the results of their experiments, analysed them, and submitted their findings to the judges. We should get feedback from the judges before the end of July…
The teams got some very interesting results!
Team Cygnus X-1 was able to capture well over 2,000 images of Earth and successfully identify and categorise photos of clouds, using a machine-learning model that they trained themselves:
The students were then able to do some post experiment analysis, back on Earth, to determine where clouds were most likely to form. They came to the following conclusions:
- Nimbostratus & cirrus clouds formed mainly over the sea
- Stratocumulus & cumulonimbus formed mostly over flat land
- Nimbostratus was the most common cloud found in warm climates
Our second team, Team CHEDDAR, took on a very different challenge – monitoring conditions within the ISS and feeding back to the astronauts if they were “wasting electricity” by leaving the lights on or having the heating too high.
They determined that environmental conditions within the space station were very consistent, with only small fluctuations in temperature and humidity. More interestingly, they were able to determine that during the three and a half hours their experiment ran, motion was detected for a total of 95 minutes. They had hoped to then compare this to their monitoring of light levels, to determine if the lights were being “left on” when not needed, but unfortunately, their experiment failed to collect information on light levels within the ISS, and so they couldn’t form a conclusion. We think that the light sensor on the AstroPi may have been covered, which meant no light was detected.
Team PASTROS carried out an experiment where they tried to identify pollution and haze clouds in the air and link them to geographical locations. Although they successfully captured photos of the Earth, upon reviewing their data it became apparent that the AI model had struggled to identify pollution and had often confused cloud cover for pollution.
They invested some time manually reviewing their photos back om Earth, but still struggled to visually identify pollution in the atmosphere.
Head of Computer Science at St John’s, Mr Jarvis said “A huge congratulations to all the students involved (Sasha Baker, Natalija Jancic-Groom, Amy Greenaway, Verity Harrigan-James, Finlay Moffat Gonzalez, Owen Mitchell and Rupert Gardner). They have shown real perseverance and creativity in their investigations and put in a lot of hours coding and reviewing their projects”.
The team members will soon be receiving a CREST Bronze Award, rewarding them for all the effort they’ve put into their projects.