AI4Mars
The AI4Mars citizen science project needs your help to teach Mars rovers how to classify Martian terrain. The rovers that have been sent to Mars have sometimes got stuck in sand as they have not recognised the change in terrain. There are a range of images that need to be identified to help with machine learning so future missions will not have these problems.
Mars is a fascinating place that has never got me bored. Join us and explore the red planet together!
Hiro Ono, AI4Mars researcher
URL : www.zooniverse.org/projects/hiro-ono/ai4mars
Reach : Worldwide
Robotic arm of the Perseverance rover
The robotic arm of NASA’s Perseverance rover is visible in this image used by the AI4Mars project. Users outline and identify different rock and landscape features to help train an artificial intelligence algorithm that will help improve the capabilities of Mars rovers.
Nature of science focus focus: Online citizen science (OCS) projects can be used to develop any of the Nature of Science (NoS) substrands. Identify aspects of NoS that your students need to get better at or understand more fully and then frame your unit to be very clear about these things when you do them.
Science capability focus: focus focus: Use evidence, Critique evidence, Interpret representations, Engage with science
Science focus focus: Astronomical systems, characteristics of technology, nature of science
Some suggested science and technology concepts:
Innovative creators of digital solutions have an impact on discovering new knowledge and understandings.
Exploring different environments.
Building curiosity and developing scientific thinking.
Importance of observation in developing effective technology.
Many concepts could be learned – focusing on a few can often be more powerful. Develop your learning outcomes and success criteria from these concepts as well as the Nature of Science strand and the science capabilities.
Some examples of learning outcomes:
Students can:
investigate that, as humans enter data for machines to learn, this can lead to bias
understand the value of machine learning for exploring the Solar System and planets
explore issues with uncrewed space exploration
develop curiosity and the wonder of discovery
understand the importance of new technology to answer scientific questions.
About the AI4Mars project
Being able to recognise the different terrain on Mars is important to successfully moving around the Red Planet. Previous Martian rovers have got into difficulties after getting stuck in soft sand. For example, in 2009, Spirit got stuck in a sand pit, which ended its mission after 7 years of exploring Mars (though this far exceeded its nominal mission length of 90 days). Both Opportunity and Curiosity also got stuck in sand but fortunately they were able to free themselves and continue on their missions.
To solve this issue, a team at NASA Jet Propulsion Laboratory is working on using machine learning – essentially the same technology used by self-driving cars on Earth. To do this successfully, the rover needs training data to learn from so it can correctly identify dangerous terrain by itself.
AI4Mars wants citizen scientists’ help in labelling a set of images captured by Mars rovers so that they can create the Solar System’s first public benchmark for Martian terrain classification. Ongoing uncrewed space exploration will depend on the rover knowing where it’s safe to drive, land, sleep and hibernate – this project is an early step in that direction.
Select a task from Perseverance – either the Geology MastCam or the Geology NavCam images.
Using AI4Mars images
AI4Mars uses images taken by Perseverance rover and wants you to draw polygons to label six geological classes: bedrock, float rock(s), sand, pebbles, peak/hill and vein.
Under the ‘Classify’ tab, there is further information on the task, with helpful video tutorials also available. Users are asked to draw polygons to label six geological classes – there is information on how to identify them and how to draw on the image.
Use the ‘Talk’ tab on the Zooniverse project website to connect to the scientists – you can ask for more information or help. Accessibility to experts has been shown to be a powerful connection and motivator with students and the subject material they are covering. Knowing that they are making a real contribution to NASA’s Mars rovers can be inspiring.
This project has been designed for year 5 up – students might need some help to start them off.
Related content
Humans have long been fascinated with Mars. Read about the Curiosity rover mission to Mars and its findings.
Future thinking is used to explore how our society and environment may be shaped in the future, including using machine learning. Start a discussion on the pros and cons of artificial intelligence using our Futures thinking toolkit . This is a useful tool to support students to develop future thinking capabilities.
See the Connected article Emotional robots . It includes activities that look at machine learning and artificial intelligence.
Observation is a key skill as explained in the article Observation and science .
Citizen science projects
The goal of the online citizen science project Planet Four is to help planetary scientists identify and measure features on the surface of Mars.
These two case studies also feature teachers who used a citizen science project in their astronomy units:
Melissa Coton (year 5/6) – measuring light pollution at night as part of a unit on light.
Matt Boucher (year 7/8) – hunting for exoplanets within a unit on light.
Here are some planning tips for when you intend to use a citizen science project with your students. See these helpful webinars: Getting started with citizen science and Online citizen science .
Activity idea
Your students may like to try the activity Is anything out there? in which they work out which planets could have life.
Useful links
These two short YouTube videos from the project are a good way to demonstrate the issues the rovers had on Mars and the work being undertaken to try and solve the problem.
SPOC-Lite: Terrain Classifier for Mars Rovers – onboard terrain classifier taking monaural images as an input and outputs the probability of sand on the image
MAARS compilation (Machine Learning-based Analytics for Automated Rover Systems research)
Learn more about NASA’s Mars rover missions:
The NASA 2020 Mission page for Perseverance rover is full of information and updates about the mission and data from Mars including media and sound files.
Check out the latest lesson plans and activities from NASA relating to Mars for educators, parents and students.
Watch this short animated video that rolls Mars around to show all the major features of the Martian topography. It begins with a hemispherical view of Olympus Mons and Valles Marineris and then rolls around to reveal the Martian South Pole. While traversing Northward, it passes Hellas Basin and ends up looking down on the Martian North Pole.
Check out the large collection of citizen science resources curated in our Pinterest board .
