77 open questions
A first build carries more unknowns than it can close, and the useful response is not to close them all but to sort them by who can answer them. The 77 questions below are grouped that way: 16 go to URC by email, 22 the team can measure on a bench, 27 need a prototype or a field day, and 12 are the team’s own decisions. Each carries a meanwhile assumption, so an unanswered question never stops work.
Key takeaways
- The judge questions go out first, in one numbered email in November 2026, because their answers take longest and the public Q&A page has changed twice in a month after publication.
- A week of measurement usually settles a decision before an answer arrives. The 22 measurable questions are cheaper than the email that asks about them.
- Every question has a meanwhile assumption, and the assumption is the conservative one. An unanswered question moves a decision to the fallback; it does not stop the season.
- Five documents are unpublished as of 2026-10-25: the 2027 review rubric, the Science Plan format, the Delivery sequence and its stage gates, the attendance numbers and the partial-credit scale for route finding.
Four groups, four clocks
| Group | Count | Who answers | What it costs to get an answer |
|---|---|---|---|
| Ask URC | 16 | The URC Director, with the answer mirrored onto the public Q&A page | A written letter and weeks of waiting |
| Measure it | 22 | The team, on a bench or with a load cell | One to three days per question |
| Test on hardware | 27 | The team, with a prototype or an analogue site | A field day and a written pass criterion |
| Team decision | 12 | The team, in writing and dated | One meeting, and the real risk is not deciding |
The 77 questions run on different clocks, which is the point of the grouping. The judge questions go out first, the measurements run in parallel, the hardware tests need a calendar slot, and the team decisions need a date and a signature.
All 77 questions
The table is searchable and sortable, and it filters by group. Each row carries the decision the question blocks, what settles it, who owns it and the meanwhile assumption.
The sixteen questions for URC
These go as one numbered letter in November 2026, each quoting its clause and stating the interpretation the team intends. Answers bind every team once they reach the public page.
| # | Question | What it decides | Meanwhile |
|---|---|---|---|
| J1 | Does one instrument with two readouts satisfy both required life-detection capabilities, or must they be independent? | One instrument or two; about 1 kg and a second fluidic channel | Two capabilities on a shared detector |
| J2 | Do “any and all batteries” include a spare drone battery carried but not connected S-URC27-RULES §2.a.v? | Battery architecture and the drone spare-pack plan | Yes; every carried pack sits behind the kill path |
| J3 | Do the antenna height and base-distance limits apply to a separate GNSS base antenna? | The RTK geometry and the 1 m stop | The limits apply to all base antennas |
| J4 | Is the Equipment Servicing cache the team’s own Astrobiology cache, or a container provided by the judges? | The arm interface and whether one cache design serves both | One cache design serves both |
| J5 | Is there a points penalty for exceeding 70 kg, or is the rover simply not fieldable? | Whether 70 kg is a trade or a gate | A gate |
| J6 | Does the oversize penalty apply per event or once? | Whether an oversize rover can still compete meaningfully | Sized to fit; the penalty is not planned as absorbable |
| J7 | Is a non-camera gesture sensor, such as an inertial unit on a sleeve, compliant with visual gesture recognition? | The 15-point Follow path | No; vision is built and audio is the reliable 10 |
| J8 | Does a repeater or drone relay need its own E-stop as a mobile deployable, or does a fuse suffice? | Relay mass and inspection | A fuse is assumed to suffice |
| J9 | Is an astronaut giving tactical directions to the C2 station “human assistance” during autonomy? | Operator discipline and whether voice may be used tactically | Voice for safety and situation reports only |
| J10 | If the astronaut presses the rover’s E-stop, is the mission terminated and are earned points kept? | The approach protocol and the stand-off | The plan keeps it from happening |
| J11 | Does the suit camera and microphone award require operators to act on the feed, or only that it is shown? | Suit electronics and the C2 layout | The feed is always shown and the judge is told it is live |
| J12 | Is there a published partial-credit scale for route finding? | Whether to attempt the far target or bank the near one | Partial credit is assumed small |
| J13 | May the 3 to 6 letter launch key contain digits? | The typing character set | Letters and digits are both supported |
| J14 | Does a compressor or vacuum pump need a Mars-feasibility exception? | Sample handling and pneumatic options | Compressors are avoided |
| J15 | What are the penalties for a sub-band violation, a late communications package or a no-spill breach? | Whether these are gates or priced risks | All three are treated as gates |
| J16 | May the science debrief pair watch live field video, or does that count as seeing the site? | Two C2-only people kept away from imagery | Two people see no imagery until the debrief |
The measurements worth starting first
The 22 measurable questions are the cheapest in the set, and six of them unblock more than one subsystem. Each is a day’s work with a load cell, a stopwatch or a rig.
- The lock: insertion force, turning torque and key free play on a pin-tumbler lock of the same class. It sets the end-effector torque budget.
- The keyboard: actuation force and travel per key, and the lateral repeatability needed at the camera standoff. It decides whether open-loop typing is possible at all.
- Backlash under load at 38 °C: the datasheet figure is nominal, and every precision claim in the arm design rests on it.
- Pack sag at the real peak: a 150 A pulse for the legitimate peak and a 200 A pulse for the trip, at low state of charge.
- Setup time: ten setups timed with the mast and 25 m of cable. The 15-minute rule is a measured quantity or an assumption S-URC27-RULES §1.a.ii.
- The cache: tube dimensions, opening torque and a gloved removal, which together decide the 5-minute hand-over.
What a hardware test looks like
The 27 hardware questions are written as pass criteria before the test runs, which is what separates them from demonstrations. Six of them carry most of the risk.
| Test | Pass criterion |
|---|---|
| Link at 1 km in a terrain shadow, on the lower band alone | Drivable at the worst shadow point; recovery inside 10 s |
| 1 m stop repeatability | 90 % of runs inside 1.0 m, and 100 % inside 1.5 m |
| Follow stand-off at the astronaut’s walking speed | 2.5 to 4.5 m for 90 % of the run, and never closer than 1.5 m |
| The bioassay inside the mission window | Both results by T+23:00, with controls behaving |
| E-stop removes all power, repeatedly | Zero volts at every rail, 50 times out of 50 |
| The packaged rover shaken for 30 minutes | Every peripheral works afterwards |
The twelve decisions
These are not facts to discover but choices, and the risk is not taking them. Each is written down, dated and signed, because an undecided scope is the mechanism behind the over-scope risk.
| # | Decision | Recommended call |
|---|---|---|
| D1 | Which of the four new 2027 capabilities the team attempts, and which it defers | Bioassay yes, because it is mandatory; route finding yes; astronaut interaction partial; key and lock last |
| D2 | One configuration for all four missions, or modular configurations | One configuration if mass allows, which removes the 10-minute re-weigh |
| D3 | Drone: fly it or leave it home | Flying only with a dedicated pilot and spotter pair by January; otherwise it stays home |
| D4 | Autonomy command modality: device, sign, audio or gesture | Device and audio; gesture as an experiment that can fail safely |
| D5 | Two independent life-detection methods, or one instrument with two channels | Two methods on one detector, so one survives the other’s failure |
| D6 | The freeze date and the known-good tag | 2027-04-16; after it, fixes only |
| D7 | The six C2 seats, the commander, and who never touches a control | The commander never drives; two operators per station |
| D8 | Whether to field the bioassay if the answer to J1 requires two independent methods and mass is binding | Mass is deleted elsewhere to fit it, not the bioassay |
| D9 | The autonomy abandonment minute and the split between sub-missions | 20 minutes or fewer for the astronaut tasks, the remainder for route finding, with a hard abandon minute |
| D10 | Whether the key and lock task is attempted at all | Cheap tasks banked first; the key attempted only if the measurements pass by week 13 |
| D11 | Which field sites the team uses, given that testing at the site is prohibited | Two legal sites by week 8 |
| D12 | Whether the team keeps and publishes its own failure log | Yes, from week 1, one paragraph per failure, with the mechanism |
What an unanswered question costs
Nothing, if the meanwhile assumption is written first. The cost appears when a team argues the same question twice, or when two rows are built on opposite guesses. That is why the register records the assumption against the affected problem rows and why the letter asks for every answer to be mirrored onto the public page: one answer, filed once, used everywhere.
The five unpublished documents are the exceptions worth watching. A missing review rubric cannot be guessed, only proxied against the 2026 one and labelled as a proxy. A missing Delivery sequence cannot be rehearsed, only ordered by points per minute when it arrives. Both are re-read monthly from December.
Further reading
- 70 risks shows which rows each answer moves.
- 110 engineering problems carries the rows these questions gate.
- The first 90 days places the letter and the measurements on a calendar.
- The Q&A page collects the rulings that already bind every team.
- 155 engineering requirements is where a question usually begins, because a requirement with no clause behind it is an open question in disguise.
Sources for this page
- S-URC27-RULESUniversity Rover Challenge 2027 – Requirements and GuidelinesOriginal Archived copy