🌱 Barkly Plant Care

Functional Requirements Document

Human-Centered Autonomous Plant Care System

Document Status: Draft Organization: Barkly Labs System: Barkly Plant Care Related Systems: PlantWater-API, CYN-X, Barkly Safety Layer

01. Purpose

Barkly Plant Care SHALL provide an intelligent plant-care system capable of monitoring individual plants, understanding their care requirements, evaluating their current environment, and adjusting care actions accordingly.

The system SHALL move unnecessary plant-care complexity away from the human while preserving human visibility, control, and override capability.

The system SHALL combine:

PlantWater-API for plant-care hardware and sensor integration

CYN-X for contextual reasoning and decision support

Barkly Safety Layer for deterministic limits and physical-system protection

Plant Profiles for species- and individual-specific care requirements

Historical Data for understanding how an individual plant behaves over time

02. Core Principle

Barkly Plant Care SHALL NOT operate primarily from fixed schedules.

The system SHOULD make decisions based on:

What plant is this? + What does this plant need? + What is happening right now? + What has happened previously?

For example:

Traditional System

Every 3 days

↓

Water plant

Barkly Plant Care:

Plant Profile

+

Current Sensors

+

Environmental Conditions

+

Historical Behavior

↓

CYN-X Evaluation

↓

Safety Validation

↓

Care Action

↓

Observe Result

↓

Update Plant History

03. System Architecture

Barkly Plant Care SHALL consist of the following logical components:

04. Plant Profiles

FR-PLANT-001 — Plant Identity

The system SHALL maintain a unique identity for every managed plant.

Each plant SHALL have:

plant_id

name

species

location

container

soil_profile

growth_stage

FR-PLANT-002 — Species Requirements

The system SHALL support storing care requirements associated with a plant species.

Requirements MAY include:

preferred soil moisture

acceptable moisture range

watering behavior

light requirements

temperature range

humidity preferences

growth stage

expected drying behavior

FR-PLANT-003 — Individual Plant Characteristics

The system SHOULD support characteristics specific to an individual plant.

Examples:

container size

soil composition

pot drainage

observed drying rate

historical watering volume

growth behavior

The system SHALL NOT assume that every plant of the same species behaves identically.

05. Environmental Monitoring

FR-SENSOR-001 — Soil Moisture

The system SHALL support measurement of soil moisture.

FR-SENSOR-002 — Temperature

The system SHALL support environmental temperature measurement.

FR-SENSOR-003 — Humidity

The system SHALL support environmental humidity measurement.

FR-SENSOR-004 — Light

The system SHOULD support measurement of available light.

FR-SENSOR-005 — Water Reservoir

The system SHALL support monitoring of available water when connected to an automated reservoir.

FR-SENSOR-006 — Sensor Health

The system SHALL detect unavailable, invalid, stale, or abnormal sensor readings.

06. Plant State

The system SHALL maintain a current representation of each plant's observed state.

Example:

Plant State

├── soil_moisture

├── temperature

├── humidity

├── light

├── reservoir_level

├── last_watered

├── recent_water_volume

├── drying_rate

└── sensor_health

FR-STATE-001

The system SHALL update plant state when new sensor data becomes available.

FR-STATE-002

The system SHALL preserve historical observations.

FR-STATE-003

The system SHALL distinguish between:

Observed

Estimated

Inferred

Unknown

The system SHALL NOT present an inference as a direct measurement.

07. Intelligent Care Evaluation

FR-AI-001 — Contextual Evaluation

CYN-X SHALL be capable of evaluating plant state in relation to the plant's profile.

The evaluation MAY consider:

current moisture

recent moisture trend

drying rate

recent watering

environmental conditions

plant requirements

growth stage

historical behavior

FR-AI-002 — Care Recommendation

CYN-X SHALL be capable of producing a structured care recommendation.

Example:

Recommendation:

action: water

suggested_volume: 80 mL

confidence: moderate

reason:

soil moisture has fallen below

the plant's preferred range

FR-AI-003 — No-Action Decision

The system SHALL be capable of determining that no action is currently required.

Example:

ACTION: NONE

Reason:

Plant moisture remains within

its preferred operating range.

FR-AI-004 — Uncertainty

CYN-X SHALL be capable of identifying insufficient information.

Example:

STATUS: UNCERTAIN

Reason:

Soil sensor has not reported valid

data for 14 hours.

The system SHALL NOT be required to make an automated care action when insufficient information exists.

08. Adaptive Plant Care

FR-ADAPT-001 — Individual Behavior

The system SHOULD learn from historical observations how an individual plant behaves.

For example:

Average drying rate:

4.2% / day

Recent drying rate:

6.1% / day

This information MAY influence future care evaluation.

FR-ADAPT-002 — Environmental Changes

The system SHOULD account for environmental changes.

Examples:

Higher temperature

Lower humidity

Increased sunlight

Different season

Different location

FR-ADAPT-003 — Growth Stage

The system SHOULD support different care profiles for different growth stages.

Example:

Seedling

↓

Vegetative

↓

Flowering

↓

Mature

09. Automated Watering

FR-WATER-001

The system SHALL support automated watering when compatible hardware is connected.

FR-WATER-002

Watering actions SHALL be generated from an evaluated plant state rather than solely from a fixed timer.

FR-WATER-003

The system SHALL support configurable maximum watering volumes.

FR-WATER-004

The system SHALL support configurable minimum intervals between watering events.

FR-WATER-005

The system SHALL record every automated watering event.

Example:

2026-09-29 11:42

Plant: Tomato-01

Volume: 80 mL

Trigger: CYN-X recommendation

Safety: Approved

10. Barkly Safety Layer

CYN-X SHALL NOT directly control physical actuators without safety validation.

Architecture:

CYN-X

│

▼

Care Recommendation

│

▼

Barkly Safety Controller

│

┌───────┼────────┐

▼ ▼ ▼

Volume Cooldown Sensors

Limit Limit Validity

│ │ │

└───────┼────────┘

▼

APPROVE / DENY

│

▼

PlantWater-API

│

▼

Pump

FR-SAFE-001

The Safety Layer SHALL enforce maximum watering volume.

FR-SAFE-002

The Safety Layer SHALL enforce minimum watering intervals.

FR-SAFE-003

The Safety Layer SHALL reject watering when required sensor information is invalid or unavailable.

FR-SAFE-004

The system SHALL provide a physical or software emergency stop where supported by the hardware.

FR-SAFE-005

A failed AI component SHALL NOT automatically bypass physical safety limits.

11. Human Control

Barkly Plant Care SHALL remain human-controlled even when automation is enabled.

FR-HUMAN-001

The human SHALL be able to view the current state of every plant.

FR-HUMAN-002

The human SHALL be able to manually water a plant.

FR-HUMAN-003

The human SHALL be able to disable automated watering.

FR-HUMAN-004

The human SHALL be able to modify plant profiles.

FR-HUMAN-005

The human SHALL be able to override an automated recommendation.

12. Explainability

The system SHALL explain significant automated actions in human-readable language.

Instead of:

WATER = TRUE

the system SHOULD display:

🌱 Tomato-01 needs water. Soil moisture has fallen below its preferred range and the plant has been drying faster than its recent average. Barkly recommends 80 mL.

The human SHOULD be able to inspect the information used to make the decision.

13. Notifications

The system SHOULD notify the human when intervention is needed.

Examples:

Normal

🌱 Fern is doing fine.

Attention

🌱 Tomato-01 is drying faster than usual.

Sensor Problem

⚠️ Tomato-01's moisture sensor hasn't reported valid data.

Hardware Problem

🚨 PlantWater reservoir is empty.

Human Decision Required

🐾 Barkly can't confidently determine whether this plant should be watered.

14. Historical Records

The system SHALL maintain a history of:

sensor observations

watering events

manual interventions

automated decisions

rejected actions

sensor failures

environmental conditions

plant-profile changes

This history SHOULD be exportable.

15. API Requirements

PlantWater-API SHALL expose machine-readable interfaces for:

GET /plants

GET /plants/{id}

POST /plants

PATCH /plants/{id}

GET /plants/{id}/state

GET /plants/{id}/history

GET /plants/{id}/sensors

GET /plants/{id}/recommendations

POST /plants/{id}/water

POST /plants/{id}/automation/enable

POST /plants/{id}/automation/disable

Exact endpoint naming MAY change during implementation.

16. CYN-X Interface

CYN-X SHOULD receive structured plant information rather than raw unorganized sensor streams.

Example:

{

"plant": {

"species": "example",

"growth_stage": "vegetative"

},

"environment": {

"soil_moisture": 31,

"temperature": 74,

"humidity": 48,

"light": 820

},

"history": {

"hours_since_watering": 19,

"recent_drying_rate": 0.061

}

}

CYN-X SHOULD return structured decisions rather than only natural-language text.

17. Failure Handling

The system SHALL fail safely.

If:

sensor fails

↓

don't guess

If:

CYN-X unavailable

↓

safe fallback behavior

If:

pump reports failure

↓

stop automated watering

↓

notify human

If:

reservoir empty

↓

do not run pump

↓

notify human

18. Privacy

Plant data SHOULD remain local where practical.

The system SHOULD NOT require cloud AI for basic plant-care functionality.

Where CYN-X is locally available, plant data SHOULD be processed locally unless the human explicitly enables external services.

19. Accessibility / Human-Centered Requirements

Barkly Plant Care SHALL follow the Barkly human-centered design principles.

The interface SHOULD:

minimize unnecessary decisions

use understandable language

expose complexity progressively

avoid requiring constant monitoring

clearly distinguish warnings from informational messages

provide manual control

provide explanations for automated actions

avoid overwhelming the user with raw sensor data

The system's purpose is not to make the human become a plant scientist.

It is to make the computer handle the complicated parts.

20. MVP

The first implementation SHOULD focus on:

MVP

│

├── Plant profiles

├── Soil moisture sensor

├── Temperature/humidity

├── PlantWater-API

├── Automated pump

├── Safety controller

├── CYN-X evaluation

├── Watering history

├── Human dashboard

└── Manual override

The MVP does not need to solve every plant-care problem.

The first goal is simply:

Identify what plant we are caring for, understand its current conditions, determine whether it needs water, safely act when appropriate, and explain what happened. 🌱🐾

21. Future Capabilities

Future versions MAY add:

plant image recognition

disease/deficiency detection

camera-based growth tracking

automatic plant identification

fertilizer management

light-control integration

greenhouse control

multiple irrigation zones

weather integration

seasonal adaptation

plant-to-plant comparisons

predictive watering

computer vision

CYN-X vision models

voice interaction

Barkly Linux/PawBoard integration

22. Barkly Plant Care Principle

The human should care about the plant. The computer should care about the complexity. 🌱💜

=