GuardKnob - Stove Safety System

Overview

GuardKnob is a comprehensive IoT safety system designed to prevent gas leaks and stove-related accidents in homes. The system combines hardware innovation with modern web technologies to create a seamless safety solution.

The Problem

Gas leaks and unattended stoves are serious safety hazards in households. Traditional stove knobs lack intelligence and safety features, putting families at risk of gas poisoning, fires, and accidents.

The Solution

GuardKnob replaces traditional stove knobs with smart, connected devices that:

• Monitor gas levels in real-time using MQ5 sensors
• Automatically shut off the stove when dangerous gas levels are detected
• Track how long the stove has been left unattended
• Provide remote monitoring and control through a web dashboard

Technical Implementation

Hardware: The system uses ESP32 microcontrollers for both the stove knob and wireless gas detector. A custom universal adapter allows the smart knob to fit on various stove models. The potentiometer maintains precise position tracking during both manual adjustment and automated shutoff.

Software: Built a ReactJS dashboard with Tailwind CSS for the frontend, providing a clean, responsive interface. The Node.js backend handles real-time communication between the hardware and web interface, enabling instant notifications and remote control capabilities.

Safety Features: The fail-safe mechanical design ensures the stove can always be controlled manually, even if the electronics fail. Multiple redundancy layers prevent false positives while maintaining rapid response to genuine threats.

Impact

GuardKnob demonstrates how IoT technology can be applied to solve real safety problems. The system provides peace of mind for families and could prevent countless accidents related to gas leaks and unattended stoves.

Future Enhancements

• Mobile app for iOS and Android
• Integration with smart home systems (Alexa, Google Home)
• Machine learning for usage pattern analysis
• Multi-stove support for commercial kitchens

Smart Trashcan Monitoring System

Award

🏆 1st Place - ECE Department Competition

Overview

The Smart Trashcan Monitoring System revolutionizes waste management on campus by providing real-time monitoring of trash levels, optimizing janitorial routes, and preventing overflow situations.

The Problem

Campus janitorial staff often waste time checking empty trash cans or miss overflowing ones. This inefficiency leads to wasted labor hours, overflow issues, and an unpleasant campus environment.

The Solution

A network of IoT sensors deployed across campus that:

• Continuously monitor trash levels using ultrasonic sensors
• Report fill levels to a centralized dashboard
• Alert staff when bins reach capacity
• Optimize collection routes based on actual need

Technical Implementation

Hardware: Each unit contains an ESP32-S2 Mini microcontroller powered by a LiPo battery. Ultrasonic sensors measure the distance to the trash surface, calculating fill percentage. The entire assembly fits in a custom 3D-printed enclosure designed to withstand outdoor conditions.

Power Optimization: Through multiple iterations, achieved a 550% improvement in power consumption. The device now has a projected battery life of 1.7 years, making it practical for campus-wide deployment without frequent battery replacements.

Software Architecture: Built a Flask web application hosted on AWS that aggregates data from all sensors. The dashboard displays a campus map with color-coded indicators showing fill levels. Staff can view historical data and receive alerts via the web interface.

Scalability

The system architecture supports hundreds of sensors, making it viable for full campus deployment. Each sensor communicates via WiFi, with fallback mechanisms for connectivity issues.

Impact

• Reduces unnecessary trips to check empty bins
• Prevents overflow situations
• Provides data for optimizing bin placement
• Improves overall campus cleanliness
• Reduces operational costs through efficiency gains

Future Development

• Machine learning for predictive maintenance
• Solar charging for extended battery life
• Integration with campus facilities management systems
• Expansion to recycling and compost bins

AI Calendar Optimizer

Overview

The AI Calendar Optimizer is an intelligent scheduling assistant that aggregates tasks from multiple sources and uses AI to help students manage their time more effectively.

The Problem

Students juggle assignments from Canvas, personal events from Google Calendar, and various commitments scattered across different platforms. Managing these effectively while prioritizing tasks is mentally exhausting and error-prone.

The Solution

An AI-powered system that:

• Automatically pulls assignments from Canvas
• Syncs with Google Calendar events
• Accepts natural language input for new tasks
• Uses Google Gemini API to intelligently parse and organize tasks
• Suggests optimal scheduling based on deadlines and priorities

Technical Implementation

Backend: Built with Flask, providing RESTful endpoints for frontend communication. MongoDB stores user data, tasks, and preferences with flexible schema design for varying task types.

API Integration: Implemented OAuth 2.0 authentication for secure Google Calendar access. Developed custom Canvas API wrapper to extract course assignments, deadlines, and grades. Both integrations handle rate limiting and error recovery gracefully.

AI Processing: Google Gemini API analyzes user input, extracting task details, deadlines, priorities, and dependencies. The AI understands context like "study for the exam next Tuesday" and converts it to structured task data.

Key Features

• Unified Dashboard: Single view of all commitments across platforms
• Natural Language Input: Add tasks by typing or speaking naturally
• Smart Scheduling: AI suggests when to work on each task
• Deadline Awareness: Automatic prioritization based on due dates
• Conflict Detection: Identifies scheduling conflicts before they happen

Data Flow

1. User connects Google Calendar and Canvas accounts
2. System syncs existing events and assignments
3. User adds new tasks via natural language
4. Gemini API processes and structures the input
5. MongoDB stores all tasks with metadata
6. AI analyzes schedule and suggests optimizations
7. User receives organized, prioritized task list

Security & Privacy

OAuth 2.0 ensures secure authentication without storing user credentials. All API tokens are encrypted and stored securely. Users maintain full control over data access and can revoke permissions at any time.

Future Enhancements

• Machine learning to learn user preferences over time
• Integration with more platforms (Notion, Todoist, etc.)
• Smart notifications via SMS or push
• Study time estimation based on course difficulty
• Collaborative features for group projects
• Analytics dashboard showing productivity patterns

Firefighting Robot

Overview

A hands-on robotics project (Nov 2024 - Dec 2024) to design, build, and program an autonomous vehicle capable of navigating to a fire, detecting its presence, and extinguishing it without human intervention.

The Challenge

The goal was to create a self-contained system from scratch. This required integrating multiple electromechanical components—flame sensors, motors for movement, a servo for aiming, and a water pump—and programming them all to work together reliably on a single Arduino microcontroller.

Technical Implementation

The robot's 'brain' is an Arduino microcontroller programmed in C++. It processes real-time data from flame sensors to detect a fire's presence. Once a fire is identified, the Arduino coordinates a series of actions:

• It maneuvers the robot into position using DC gear motors.
• It aims a water nozzle precisely at the flame using a servo motor.
• It activates a high-voltage water pump to extinguish the fire.

This project required a complete end-to-end design, from soldering and circuit assembly to programming the control logic for all hardware components.