University students working with 3D-printed rovers as part of an ESA Academy training course at ESA’s ESTEC technical centre in the Netherlands last week. Six student teams were tasked with enabling their rovers to autonomously search and navigate towards a blue ball within a simulated Mars environment.

Each of the teams received their own ExoMy rover – a fully 3D-printed rover inspired by ESA’s Rosalind Franklin design, with six wheels, a camera and Raspberry Pi computer. Both ExoMy’s hardware and software are fully open source, along with extensive building and assembly instructions.

“This ESA Academy Robotics Workshop 2024 involved a total of 30 university students with an engineering or robotics background, from 14 different ESA Member States and Canada,” explains robotics engineer Marti Vilella Ramisa of ESA’s Automation Robotics section.

“The participants did not need to have previous familiarity with the topics involved; the objective was to familiarise them with the design and operation of a 3D-printed rover, inspired by ESA’s Rosalind Franklin ExoMars rover scheduled to be sent to investigate the surface of the red planet by the end of this decade.

“The four-day workshop involved a mixture of lectures and tutorials then hands-on exercises to put their new-found knowledge to the test. This involved adding features and fixing bugs in the rovers’ ROS 2 Robot Operating System, mostly programmed in Python.”

The workshop began with a general introduction to robotics and its application to Mars exploration. Next, they were introduced to the ExoMy rover and the code behind it. With a wireless gamepad, they tried out driving a simulated version within a computer application as well as the physical rover across a Mars-like surface.

Next the teams learned about the various locomotion modes used in rovers and the algorithms behind them, using this knowledge to implement two new locomotion modes on their ExoMy, increasing their options for operating on the simulated Mars-scape.

Each student team then used the camera on their ExoMy to take hundreds of photos of the target ball on the Martian surface. By labelling these images with the object location they could use a Machine Learning algorithm to train a neural network to recognise the ball by itself.

“The final exercise involved combing locomotion and image detection capabilities to find and drive towards the ball, allowing for the fact that the ball would be located at an unknown point,” adds Marti. “This was a challenging task, but all the teams proved successful, and some were over and above their trainers’ expectations!”

The teams presented their solutions and were graded accordingly, receiving a course transcript that, along with their certificate of participation, the students can ask for European Credit Transfer and Accumulation System credits at their universities.

More ESA Academy training courses and opportunities are coming: find out more here. For inquiries contact .