04 Intelligent Modules
81% Evaluation Weight Covered
AI + IoT Integrated Architecture

Project Introduction

SmartRose is an AI-IoT research platform for greenhouse rose cultivation, designed to support data-driven and timely decisions from growth to post-harvest handling.

Outcomes & deliverables

Problem definition, architecture, models, and experiments feed into reports, presentations, the research paper, and the published portfolio.

Close-up roses inside greenhouse Greenhouse rows with rose plants Wide greenhouse with roses

Research

Domain

Survey, gaps, and where SmartRose sits in the rose lifecycle.

Project scope

Literature Survey

Controlled-environment agriculture increasingly combines IoT sensing with ML, but end-to-end rose lifecycle coverage remains limited in published work [1].

High-level SmartRose architecture: IoT sensing units, ESP32 controller, database, trained ML prediction modules, and mobile app for farmers and florists.
Figure 1. Indicative trend supporting the survey context [1].

AI-IoT adoption in controlled agriculture has shown strong potential for predictive, data-driven systems. Most prior work remains task-specific and does not cover the complete rose cultivation lifecycle from propagation to post-harvest shelf life.

SmartRose situates itself in this landscape by targeting an integrated platform: risk prediction, nutrient intelligence, stress-aware operation, and freshness estimation—areas that are rarely unified in a single research and engineering effort.

References

  1. Full survey citations appear in the research report and follow the program referencing standard.

Research Gap

Three limitations in current work motivate SmartRose: fragmentation, reactive control, and weak post-harvest linkage.

Task fragmentation

Existing systems often address single tasks (e.g. irrigation or a single model) without an integrated lifecycle view for greenhouse roses.

Reactive decisions

Many pipelines still depend on visible stress or manual checks, reducing the value of early warning and proactive control.

Information & post-harvest gap

Freshness and vase-life signals are seldom connected back to cultivation-stage models for decision support.

Research Problem & Solution

How can greenhouse rose cultivation be optimized end-to-end using a unified platform that supports early warning, nutrient intelligence, stress-aware control, and post-harvest prediction?

SmartRose addresses this through a single architecture that combines sensing, machine learning, and decision support for growers and floriculture stakeholders.

Architecture

Core Intelligent Modules

Four ML features: risk, nutrients, environment, and post-harvest freshness.

01

Early Disease Risk Prediction

Binary stress-risk classification using environmental and physiological sensor signals.

02

Intelligent Nutrient Management

Nutrient forecasting and fertilizer optimization with regression-based machine learning.

03

Stress and Energy Optimization

Stress-level prediction with recommendations for ventilation, cooling, lighting, and irrigation.

04

Post-Harvest Freshness Prediction

Freshness score and remaining vase-life estimation for improved quality preservation.

Approach

Methodology

Sensors, connectivity, learning models, and the grower dashboard—plus the main inputs and outputs used.

01

Sensing Layer

Captures temperature, humidity, soil moisture, pH, EC, NPK, light, and leaf temperature.

02

Communication Layer

Streams data through Wi-Fi and LoRa to support both local and distributed greenhouse setups.

03

Intelligence Layer

Runs Random Forest-based classification and regression for disease risk, stress, and freshness.

04

Application Layer

Delivers recommendations, risk alerts, and monitoring summaries through a user dashboard.

Input Feature Groups

  • Environmental: Temperature, Humidity, Light
  • Soil: Moisture, pH, EC, NPK
  • Physiological: Leaf temperature, Delta-T indicators
  • Post-harvest: Ethylene, water level, storage duration

Model Outputs

  • Healthy vs Stress-Risk classification
  • Nutrient recommendation values
  • Stress level: LOW / MEDIUM / HIGH
  • Freshness score and vase-life estimate

Stack

Technologies Used

Languages and tools for APIs, ML, data, and client apps.

Python
Python
FastAPI
FastAPI
Scikit-learn
Scikit-learn
NumPy
NumPy
Pandas
Pandas
MongoDB
MongoDB
Flutter
Flutter
Dart
Dart
React
React
TypeScript
TypeScript
Tailwind CSS
Tailwind CSS
Vite
Vite
Arduino / IoT
Arduino / IoT

Milestones

Timeline in Brief

Coursework steps and mark weights; the entries below account for about 81% of the final grade.

Sep 11, 2025

Project Proposal

Initial problem definition, scope, and proposed SmartRose architecture.

Jan 7, 2026

Progress Presentation I

Approximately 50% completion, preliminary results, and validation plan.

May 8, 2026

Research Paper

Contribution, methodology, experiments, and analysis in manuscript form.

Mar 11, 2026

Progress Presentation II

Advanced integration, system validation, and near-final design review.

Apr 26, 2026

Website Assessment

Portfolio structure, content quality, and presentation of research work.

May 13, 2026

Final Report

Complete documentation, final results, and individual contributions.

May 6, 2026

Final Presentation & Viva

Final demo, Q&A, and individual oral examination.

Validation

User Testing Video

Watch the SmartRose user testing session directly on the website.

Downloads

Documents

Project documents and submission files for download.

Research Paper

PDF

25-26J-299 Research Paper Download

Group Thesis

PDF

25-26J-299 Group Thesis Download

Presentations

Slides

Progress and final presentation slide decks by review stage.

Final presentation

Pending

Group Pending

People

About Us

Supervisors and student team, with contacts.

Supervisors

Anjana Junius Vidanaralage

Anjana Junius Vidanaralage

Supervisor

Faculty

Sri Lanka Institute of Information Technology

Department
Computing

E-Mail
Kaushika Kahatapitiya

Kaushika Kahatapitiya

Co-Supervisor

Faculty

Sri Lanka Institute of Information Technology

Department
Computing

E-Mail

Our Team

Maleen Rodrigo

Maleen Rodrigo

Group member

Undergraduate

Sri Lanka Institute of Information Technology

Department
Information Technology

LinkedIn E-Mail
Savinda Rajapaksha

Savinda Rajapaksha

Group member

Undergraduate

Sri Lanka Institute of Information Technology

Department
Information Technology

LinkedIn E-Mail
Chamikara Bandara

Chamikara Bandara

Group member

Undergraduate

Sri Lanka Institute of Information Technology

Department
Information Technology

LinkedIn E-Mail
Mahimi Perera

Mahimi Perera

Group member

Undergraduate

Sri Lanka Institute of Information Technology

Department
Information Technology

LinkedIn E-Mail

Outreach

Project updates

Releases and docs as the project moves.

Reach out

Contact us

Send a message to the SmartRose project team.

^