Week 13 — Break, midterm review
Officially Week 13 is the Mid Term Review: system diagram, remaining tasks, schedule, and meetings with local and global instructors. I treated it as a checkpoint for Forest Fairy, not another week of quietly adding parts without updating the plan.
Individual assignment
The midterm review forced a pause. I had been adding subsystems faster than I was explaining how they connect. Forest Fairy is a mobile plant companion for a chili pepper plant: it reads the plant and its environment, interprets those readings cautiously, and responds through display, dialogue, and movement. This week I mapped the system, named what was still missing, and turned the remaining course weeks into a schedule I could actually follow.
1. Official review checklist
The Fab Academy midterm page asks for progress on previous assignment documentation, progress on the final project idea, a schedule for the remaining tasks, and meetings with instructors. I checked my site against those points and used the result to update the Final Project midterm review section.
I posted the system diagram on final-project.html so the main project page shows how sensing, computation, interface, and motion fit together. I split the remaining work into mechanical, electronics, sensing, output, integration, and documentation tracks instead of one vague todo pile. Weeks 14–20 now map to concrete deliverables, from molding/casting through final presentation. I also went back through Week 1–12 pages for missing photos, source links, and proof. For instructor review I kept the questions in section 6 below; I did not invent names or dates I had not confirmed.
2. Final project system diagram
The full diagram is on the Final Project page. I drew it at system level, not down to every pin, because the midterm question is whether someone else can follow the story: plant and environment signals go into the ESP32-based hub, the hub builds a plant-state object, and that state drives the UI, voice/AI layer, and motion layer.
Drawing it that way surfaced a risk I had been avoiding. If I try to perfect every sensing channel before I have an end-to-end demo, integration week will slip. So my midterm cut is to keep the core loop small first: one stable environment reading, one plant-side or soil reading, one readable screen, and one constrained movement or prompt.
3. Tasks to be completed
The diagram shows the target architecture; the table below is what still has to land on the bench. I grouped work by subsystem so I can see where a slip in one area blocks everything else.
| Area | Remaining work | Why it matters |
|---|---|---|
| Mechanical | Finish planter-to-base mounting, cable strain relief, and safe access for watering. | The robot must survive handling before any sensing story is believable. |
| Electronics | Populate/test the carrier board, confirm power rails, label I²C/UART/SPI paths, and keep a known-good fallback. | Late wiring confusion is the fastest way to lose integration time. |
| Sensing | Bring up temperature, humidity, light, soil, and plant-side acquisition with repeatable logs. | The plant dialogue should come from measured change, not decorative text. |
| Outputs | Finish display pages, voice prompts, motion limits, and one short demo sequence. | Reviewers need to see the system respond in a human-readable way. |
| Integration | Merge the firmware paths, test buses under battery power, and freeze demo behavior before final week. | The goal is one stable loop, not many isolated demos. |
| Documentation | Close missing photos, link source files, add captions, and keep failed attempts visible. | The site must let another person understand what I made and how I made it. |
4. Schedule for remaining weeks
I turned the remaining course weeks into a delivery plan with one main target per week. The same table lives on the Final Project page so I do not maintain two different versions of the schedule.
| Week | Main target | Evidence I should publish |
|---|---|---|
| Week 14 | Molding/casting details for Forest Fairy branding or soft parts. | Mold design, material notes, casting photos, and result. |
| Week 15 | Interface and application layer. | Screen layout, code, screenshots/photos, and user flow. |
| Week 16 | System integration. | Wiring, firmware merge, power test, and an end-to-end demo attempt. |
| Week 17 | Wildcard / risk reduction. | Fix the weakest subsystem instead of adding a new feature. |
| Week 18 | Applications, implications, and project context. | Use cases, audience, limitations, safety, and ethics. |
| Week 19 | Invention, IP, and income. | License choice, ownership, and possible continuation path. |
| Week 20 | Final presentation and documentation freeze. | Video, slide, final photo, final source archive, and navigation cleanup. |
5. Weekly assignment documentation review
I scanned Week 1–12 against the official midterm prompt. The trail is continuous, but several pages still read like isolated weekly exercises instead of steps toward Forest Fairy. Early weeks (framing, CAD, cutting) need clearer links back to the chassis and planter structure. Weeks on embedded programming and electronics design need source links and board/version notes, or later me cannot tell which firmware tree matches which spin. Machining, board production, and input-device weeks should show what actually landed in the final stack, not only what passed that week's assignment. Output, networking, and machine-design pages feed integration directly, so I flagged places where photos or diagrams were hard to find from the final project page.
Day-to-day tracking stays in
final-project-weekly-plan.md; this page is the
midterm summary. From Week 8 onward I used Cursor to code and refactor the firmware trees the schedule
points at (hub, WROOM display, Pico bridge, charging dock) and kept flashing each tree to hardware as I
merged paths for integration.
6. Review meetings
I did not have confirmed instructor names and dates in my notes when I wrote this page. Rather than paste placeholders, I wrote down what I actually need from local and global review.
For local review I want to know whether the system diagram makes the plant, hub, display, voice, and motion roles clear enough that someone else could explain the project without me in the room. For global review I want a sharper cut list: if integration time gets tight, which subsystem should I shrink or drop first? After either meeting I will update the task table and Week 14–20 schedule once feedback is confirmed.
I am not trying to defend every feature on the diagram. I need a build that can be finished.