Plant Whisperer

Mission

Build this

Build a four-choice interactive OLED label that presents one reviewed fact, one carefully contextualised live reading, and one courteous bow.

Bill of materials

Parts

Disconnect USB before rewiring. Motors, the relay, and the servo need appropriate power and a shared ground. Never drive an actuator from the ESP32 3V3 pin.

The central idea

One answer can contain two different truths

If the moisture probe is physically associated with only one plant, say so. For the other selections, label it as greenhouse context or omit it. A clear limitation is better than a personalised-looking number with no real connection.

Write four tiny introductions

Severe space can improve the facts

Wiring

Choose one input method and make it dependable

The reference route uses four labelled keys on an IR remote. A keypad or RC522 reader is valid, but begin from its working kit example and repeat the same thirty labelled selection trials.

The power rule: The OLED and verified IR receiver module use 3V3 logic. The servo uses a separate 5 V supply. Join all grounds.

Make one selection become one complete answer

The bow comes after the device understands the request

Your two-part evaluation

Test input reliability and understanding separately

1. Run at least thirty labelled selection trials.

   Vary distance and angle for the remote, tag, or keypad. Record wrong, missed, and duplicate selections.

2. Freeze four facts and the context wording.

   Have garden staff review accuracy, brevity, and whether live data is connected honestly.

3. Invite at least five first-time visitors.

   Ask them to select a named plant, repeat one fact, and explain what the live number describes.

4. Revise the weakest layer.

   Input failures and comprehension errors are different problems. Change the correct one, then repeat its test.

When the label whispers nonsense

Find whether the failure began in input, content, or context

Choose your interpretation problem

What kind of plant-label team will you become?

A calm way through the build

Collect four small wins

1. pw-01 says hello.

   Join the garden before writing the exhibit.

2. Four controls produce four registered IDs.

   Test input mapping before adding facts.

3. One reviewed fact and one live context screen are distinct.

   Add source, unit, location, age, and no-data language.

4. One valid selection earns one bow.

   Power the servo separately, then begin the two-part evaluation.

The garden handshake

Share what you found

These names are the rigid part of the project. They let another team find your work without knowing what you called the variables in your code.

Finish line

Ready to introduce to the garden

• □

  pw-01 stays online and publishes only plant-1, plant-2, plant-3, or plant-4 after valid selections.

• □

  The reject feed stays clear after the final code starts.

• □

  The label reads Thirsty moisture with unit, source, location, age, and honest stale or no-data states.

• □

  Each accepted selection produces one matching answer and one non-blocking bow.

• □

  The build log contains thirty labelled input trials, wrong, missed, and duplicate requests, four reviewed facts, five visitor tasks, fact recall, context comprehension, and one targeted revision.

• □

  The chosen input hardware uses safe voltage, the servo is separately powered, content is readable, and the device label can be scanned.

The Plant Whisperer is most persuasive when it knows which voice it is using: a carefully edited fact, a nearby measurement, or an honest admission that the connection is not known.