The Big Debut That Wasn’t
The ERC 2025 cycle was supposed to be our big debut. Lunomyss, the rover, and Aetheromyss, the drone, were the crown jewels of months of design reviews, sleepless nights, and more budget spreadsheets than anyone sane should tolerate. What began as sketches and half-serious “what if we made it weird” ideas hardened into a full preliminary design: active rocker suspension, all wheel steering, an onboard Mac mini brain, a swarm of comms systems, and a science payload that could chew into Martian soil and tell us what it had for breakfast. We were not building a rover to just tick boxes. We wanted something that could bend the competition field into its orbit.
“We were not building a rover to just tick boxes. We wanted something that could bend the competition field into its orbit.”
Design Battles
The design steps were brutal but intoxicating. Every subsystem was questioned, reworked, and fought over. The suspension turned into a feat of electrical adjustability, each arm controlled by a linear actuator so Lunomyss could crouch, stand tall, or stabilize for drilling. The chassis evolved from aluminum profiles into dreams of SLM printed metal, embedding cable channels and mounts directly into its skeleton. The wheels became a revolving experiment in traction, 3D printed, swapped, broken, iterated, until they felt like claws meant for Polish test fields.
The Brain
The brain was equally ambitious. A Mac mini M2, dockerized ROS stack, custom USB to serial distribution, VESC motor drivers, and the ever growing swarm of sensor inputs. Four external cameras, two on the arm, a solid state LiDAR, IMUs, weather station, and a lab on a chip for real time soil chemistry. For autonomy, Lunomyss carried a fragment of our other project: The Great Coral Reef, a miniature AI co processor tasked with handling image classification and terrain segmentation on the fly. No GPS, no ground control lifeline. Just sensors, dead reckoning, and enough code to make it think fast.
Aetheromyss
And then came Aetheromyss. Not an accessory but a full scout in its own right. A carbon fiber frame, carbon PETG armor where it mattered, outrunner motors pushing it through the air, and a Raspberry Pi Zero 2 as its brain. AI tasks ran on a baby Coral Reef, while the XR2 Nano radio kept it leashed when needed. It was meant to give Lunomyss eyes in the sky: fast terrain mapping, aerial recon, and a backup communication relay if the ground got too messy. We did not want just a rover. We wanted a system, a pair, a ground and air duo that worked like one organism.
Research Payloads
ERC 2025 was also meant to be the first real world testbed for the custom research we had been developing in parallel. A Lab on a Chip module, capable of analyzing soil samples chemically on the spot. A refined version of our Coral accelerator, designed for onboard machine learning with fault tolerance in mind. And our own inertial dead reckoning navigation system, built to survive without GPS and stitched together from IMU data and clever filtering. These were not just bolt ons. They were core research elements meant to push the frontier of what student built rovers could carry.
Reality Bites
But no project survives contact with reality without scars. The test plan caught weaknesses the design slides never admitted. Steering servos drifted instead of snapping back to center. Dust ingress mocked our environmental sealing. Vibration testing revealed connections that loved to wiggle loose at exactly the wrong time. Autonomous navigation algorithms tripped over themselves once sensors met the real world instead of a simulator. And looming over all of it was the same old nemesis: money.
The Budget Wall
The budget was tight from the start, a patchwork of personal contributions, favors, and creative stretching. We tried to keep costs low with 3D printing and off the shelf parts, but a 50 kg rover is never a cheap date. Motors, actuators, sensors, batteries, travel, everything added up. By the time we pushed into integration, we had designs ready, parts sourced, and systems on the bench. What we did not have was enough runway to finish the full build before the competition clock ran out.
“A 50 kg rover is never a cheap date. The budget refused to stretch another millimeter, and that heartbreak will haunt us into 2026.”
Scars, Not Failures
That hurt. Watching a project with so much potential stall in the hangar because the budget refused to stretch another millimeter was the kind of heartbreak every student team fears. We had the architecture. We had the testing. We had the plan. We just did not have the final machine rolling in Poland this year.
Still, ERC 2025 was not a failure. It was a rehearsal, a full scale design campaign that tested our process, revealed our blind spots, and proved that we can take on complex, multidisciplinary projects and make them real, even if not fully realized yet. Lunomyss and Aetheromyss exist, not as entries on a leaderboard, but as platforms waiting to be finished. And the custom research we poured into them, the LoC, the Coral, the inertial navigation, lives on, already carrying the Shobolinsky name into future projects.
Looking Ahead
We will be back in 2026 with the lessons baked in, the scars turned into improvements, and the same stubborn refusal to build something ordinary. The rover will roll. The drone will fly. The research will graduate from prototypes into hardware that works in the wild. And the rats will return, louder and stranger than ever.
“The rover will roll. The drone will fly. And the rats will return, louder and stranger than ever.”