From analysis to build
The earlier parts of this book end in three lists: the problems a 2027 rover creates, the ways a season can fail, and the questions nobody on a first build can answer yet. This part turns those lists into a plan. It carries 110 engineering problems in five dependency waves, 70 risks with a stated likelihood and impact, and 77 open questions grouped by who can answer them. The plan runs from 2026-10-26 to the finals on 2027-06-02, and both of its hard gates fall in the first half of that window.
Key takeaways
- The PDR is due 2026-12-02, five weeks into the season S-URC27-RULES §3.a.ii, and the SAR is due 2027-03-05, ten days before the integration and test wave opens S-URC27-RULES §3.a.iii. Both are gates that close a wave, not reviews that happen at the end.
- The first four things to do are measurements. The numbers that size the rover are not published, so the season is won by replacing them with measured values before the design is committed.
- The register is ordered by dependency, not by subsystem. Thirteen rows have five or more dependents; they decide when everything else becomes possible.
- Nine risks carry the argument, and six of the nine are about process rather than hardware: the review, the build target, integration time, depth of the team, scope and mass tracking.
Three registers and one calendar
| Part of the plan | Rows | The question it answers | Where it is set out |
|---|---|---|---|
| Engineering problem register | 110 | What has to be learned, decided or tested, in what order, and what closes each item | Five waves, 253 dependency edges, 155 traced requirements |
| Risk register | 70 | What can go wrong, how likely it is on the evidence, and what it costs if it happens | Six categories, a likelihood and impact band per row |
| Open questions | 77 | What is unknown, who can answer it, and what the team assumes until they do | Four groups: URC, measurement, hardware test, team decision |
| The first 90 days | 13 weeks | Which work unblocks the most other work, and in what order | The register’s dependencies, read forwards |
| The 13-week plan | 13 weeks | Which decision is taken in which week, and who owns the row | A dated calendar with four gates |
The registers are not a schedule and they are not a wish list. They make no component recommendation of their own and they carry no cost ranking: priority is set by rule consequence and dependency, never by price. The design recommendations live in the subsystem and trade pages.
The five waves
| Wave | Weeks | Calendar | What it is for | Gate |
|---|---|---|---|---|
| W0 Measure | 0–4 | 2026-10-26 to 11-22 | Bench, rig and field measurements that replace the rulebook’s missing numbers; a judge-like task mock | none |
| W1 Architecture | 4–12 | 2026-11-23 to 2027-01-17 | The irreversible or expensive-to-reverse choices, with a freeze date named for each | PDR 2026-12-02, week 5 |
| W2 Detailed design | 12–20 | 2027-01-18 to 03-14 | Interface documents, sizing analyses, procedures, and the February evidence runs the SAR needs | SAR 2027-03-05, week 19 |
| W3 Integration and test | 20–27 | 2027-03-15 to 05-02 | The rover assembled and the four missions run against a clock | test evidence for the finals |
| W4 Review and rehearsal | 27–31 | 2027-05-03 to 06-01 | Review responses, frozen procedures, rehearsal, and the final documents | Science Plan 2027-05-21, finals 2027-06-02 to 06-05 |
The communications documentation is due 2027-04-30, in the last week of W3 rather than in the review wave, which is a detail that changes where the work sits.
The two gates inside the first half
The PDR falls five weeks into the season, inside the architecture wave, at a point where most first builds have not finished measuring. The SAR falls ten days before the integration and test wave opens, so the rover has to be filmed at partial fidelity while it is still being assembled. The consequence is structural: the February evidence runs are a scheduled part of the detailed-design wave, not an afterthought, and a team that discovers the ordering in March has already lost the submission.
The same two dates explain the shape of the risk register. The highest-ranked risk is missing the SAR cut, where two of three registrants never reach a score sheet, and the second is building to June instead of building to March.
How to read a register row
Every problem row is read left to right, and the structure is the same for all 110 rows.
| Field | What it gives | Why it exists |
|---|---|---|
| ID | A stable identifier for the life of the project | The risk register, the reviews and the team tracker all reference it |
| Problem | One testable sentence with no solution inside it | A problem that contains its answer is a decision already taken |
| First experiment | The smallest thing that produces the missing information, sized in days | If it needs the whole rover, it belongs in a later wave |
| Success criterion | A number or a binary, written before the experiment runs | Rule numbers are quoted; a design target is labelled as derived |
| Unlocks | How many later problems depend on this one | The working measure of downstream design unlocked |
Priority is deliberately coarse: 29 Critical, 72 High, 8 Medium and 1 Low. The fine ordering is the unlock count, and the two are read together. A High row with eight dependents is better value than a Critical row with none.
The questions, in four groups
| Group | Count | Who answers | How long it takes |
|---|---|---|---|
| Questions to URC by email | 16 | The URC Director, with answers mirrored onto the public Q&A page | Weeks, and the page has changed twice in a month after publication |
| Measurements the team can take | 22 | The team, on a bench or with a load cell | Days |
| Tests that need hardware or a field day | 27 | The team, with a prototype or an analogue site | Weeks, and they need a date |
| Decisions the team takes | 12 | The team, in writing and dated | One meeting each, and the risk is not deciding |
Each question carries a meanwhile assumption, so an unanswered question never stops work. The sixteen judge questions go out first, in one numbered email in November 2026, because they take longest to come back.
The calendar in three views
| View | Span | What it is for |
|---|---|---|
| The 90 days | 2026-10-26 to the end of January 2027 | The measurement wave, the PDR, and most of the architecture decisions |
| The 13 weeks | 2026-10-26 to 2027-01-23 | The same window as a dated calendar with one decision per week and four gates |
| The season | 31 weeks to 2027-06-05 | Five waves, four review submissions and the test window after the SAR |
After the SAR there are 89 days to the finals, and they are a test window rather than a build window S-URC27-RULES §3.a.iv. That number is the reason the gate dates matter more than any individual subsystem deadline.
Where to start
- A new team, first week: read the problem register for the five waves, then the first 90 days for the order of work. The first four rows to start are a mass frame, a task mock, a link site booking and a kill-chain breadboard.
- A team that has a design: read the risk register first, and check the nine high-likelihood, high-impact rows against the current plan.
- A team in review season: read the open questions, send the sixteen judge questions, and write the meanwhile assumptions onto the affected rows.
- Following the argument in order: coupling and the trade studies supplies the recommendations that each register row exists to test.
What the plan assumes
The plan holds while five things are true: the 2027 rules stay as published, with the 50 kg and 70 kg limits and the 1.2 m cube; the team can build and measure on a bench and at two analogue sites before June; at least six people can be trained to hold the C2 seats; long-lead parts are ordered against numbers frozen in weeks 6 to 9; and the questions are answered or their fallbacks accepted by their dates.
If one of the first four fails, the dates move and the order does not. A register that is read once is a document; one that is reviewed weekly is a management system, and the problem register describes the 45-minute weekly review that keeps it current.
Further reading
- 110 engineering problems sets out every row with its experiment and criterion.
- 70 risks scores each row and names the change that removes it.
- 77 open questions lists the questions, the decision each blocks and the meanwhile assumption.
- The first 90 days reads the register’s dependencies forwards.
- The 13-week plan turns the same window into a dated calendar with four gates.
- Coupling, conflicts and twelve trade studies supplies the decisions the register tests.
Sources for this page
- S-URC27-RULESUniversity Rover Challenge 2027 – Requirements and GuidelinesOriginal Archived copy