Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

63 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

EcoBloom

Making energy visible. Making waste emotional.


✧ Inspiration

How can AI be resource-efficient while solving environmental problems?

That question led us beyond software… into something tangible.

What if energy waste wasn’t just a number on a dashboard, but something you could feel?

EcoBloom is our answer: a living, reactive system that transforms invisible energy usage into an emotional experience.

People don’t change behavior from data. They change from feeling.

A dashboard saying:

“Lights on for 23 minutes”

gets ignored.

But a plant slowly wilting because of that same action? That’s impossible to ignore.


✧ What It Does

EcoBloom observes, interprets, and responds in real time.

The Plant (Hardware)

A physical plant powered by green → yellow LED gradients that reflect your energy behavior:

  • 100 (Healthy): Fully green
  • 0 (Dying): Fully yellow
  • Smooth crossfade between states

It doesn’t just display data — it feels alive.


LLM Chatbot (AI)

Ask:

“Why are you dying?”

EcoBloom responds with real explanations:

“Lights were on for 145 seconds with no one present.”

Powered by Backboard, it uses actual sensor logs, not generic advice.


The Dashboard (Web)

A responsive interface showing:

  • Live mood / health score
  • Event history
  • Visual plant synced with hardware
  • Integrated chatbot

A digital mirror of the physical system.

⚙️ The Backend (Real-Time System)

Running on QNX (Raspberry Pi):

  • Camera-based room state detection
  • Event logging with precise timing
  • Stateful mood computation
  • API powering all components

Everything updates in real time.


✧ How We Built It

Tech Stack

  • Backend: Python Flask + QNX RTOS (GPIO control)
  • AI Layer: Backboard.io (LLM orchestration + grounding)
  • Frontend: React + HTML/CSS
  • Hardware: Raspberry Pi + LED system

Design Decisions

✦ Stateful Mood System Tracks duration, not just events → avoids false alarms, captures real waste

✦ Grounded AI Responses Backboard ensures outputs come from actual logs → no hallucinations

✦ Real-Time Guarantees (QNX) Sub-100ms LED response → makes the plant feel alive

✦ Mock-First Development Simulated camera inputs early → unblocked full system development


✧ Challenges

🔌 Hardware Reality Wiring LEDs + GPIO took longer than expected → physical systems ≠ instant feedback

🔄 System Coordination Multiple services required strict API contracts → solved with early mocking

🧠 Backboard Learning Curve Understanding routing + threads took time → but enabled powerful grounding

📸 Sensor Noise Camera flickering in low light → fixed with temporal smoothing


✧ Accomplishments

✔ Full end-to-end pipeline working ✔ AI responses grounded in real data ✔ Smooth, continuous LED transitions ✔ Deterministic real-time system (QNX) ✔ Demo-ready in 36 hours


✧ What We Learned

Real-time changes everything Latency isn’t just technical — it’s perceptual.

Hardware humbles you Every wire matters.

AI is more than prompts Orchestration + grounding = real systems.

Emotion > Information People don’t respond to numbers. They respond to life.


✧ What’s Next

🚀 Immediate

  • Finalize LED hardware wiring
  • Launch full physical demo
  • Deploy backend (Render / Railway)

🌍 Future Vision

  • Multi-room energy tracking
  • Smart home integrations
  • Personalized behavioral insights
  • Scalable sustainability platform

✧ Final Thought

EcoBloom isn’t just a project.

It’s a shift in perspective:

What if technology didn’t just inform us — but made us care?


About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages