User Needs

• Know whether a liquid is still safe to consume

• Avoid relying on smell/visual judgment (often unreliable for liquids)

• No complex operation—status should be instantly readable

• Reduce waste while preventing accidental consumption of spoiled liquids

• Build trust through consistent, explainable monitoring over time

System Flow

Liquid → Sensor-embedded Bottle (pH / Temp / TDS / Turbidity) → Microcontroller → MQTT → Data Hub / Server → Dashboard (User View) + AI Prediction Layer

Key Features

• Bottle: Integrates multi-sensor readings for real-time liquid-state monitoring

• Dashboard: Displays current sensor values + simplified status (Safe / Risk / Alert)

• MQTT Streaming: Low-latency transmission for live updates

• On-device Feedback: LED + OLED status for at-a-glance checking

• AI Add-on: Predicts spoilage risk by learning patterns from time-series trends

Strategic Thinking

• Designed a non-intrusive freshness-checking experience that fits daily routines: “check the bottle, not the data.”

• Prioritized multi-signal sensing (not a single metric) to better reflect complex liquid changes.

• Identified a key challenge: different liquids have different baseline values, making fixed thresholds unreliable.

• Shifted the logic from absolute standards to baseline + fluctuation detection, enabling the system to adapt across edible liquids.

• Framed the product as a scalable ecosystem (bottle + dashboard + prediction), extending from personal use to shared households.

Tech Stack (Prototype)

• Hardware: pH sensor, turbidity sensor, TDS sensor, DS18B20 temperature probe, LEDs, OLED screen, Arduino Nano 33 IoT

• Software: MQTT (data streaming), MQTTX (debug), Web dashboard, Node.js + MySQL (persistent logging), AI prediction layer (planned)

• Why MQTT: Lightweight, real-time IoT messaging ideal for streaming sensor telemetry