RISHABH FAN CLUB

Inspiration

While volunteering at an elderly home, we witnessed firsthand the communication barriers between caregivers and elderly residents. Many seniors had difficulty articulating their needs, emotions, or discomfort — often leading to delayed assistance or misunderstandings. This experience revealed a pressing need for a more intuitive, accessible, and empathetic way to communicate care needs.
This inspired us to design a system that empowers elderly users to express themselves effortlessly and enables caregivers to respond quickly and accurately through connected technology.

What it does

CareSync Companion is an integrated hardware–software solution that bridges the gap between caregivers and care recipients.

The wearable device allows the elderly to indicate how they feel using simple tactile buttons corresponding to emotional states (e.g., In Pain, Want Company).
The caregiver web app instantly receives these inputs, providing visual alerts, data logs, and analytics dashboards.
Beyond real-time interaction, the system’s backend analyzes event frequency, timing, and context to detect patterns — offering predictive insights about potential stress, fatigue, or inactivity.
By combining simplicity of use with intelligent data interpretation, CareSync Companion enhances trust, empathy, and efficiency in care environments.

How we built it

We developed the project through multiple layers of integration:

1. Hardware Layer

Built around ESP32 microcontrollers equipped with tactile buttons, RGB LEDs, and a buzzer.
Implemented UDP-based communication for low-latency message transfer between caregiver and care-recipient devices.

2. Backend Layer

Developed using Node.js and Firebase to handle event streams, authentication, and device state management.
Implemented Server-Sent Events (SSE) for seamless real-time updates to the web dashboard.

3. Frontend Layer

Built a React-based web application to visualize live device data, emotional trends, and risk scores.
Integrated Firebase Firestore for persistent data storage and Chart.js for interactive analytics and UI feedback.

4. Machine Learning Layer

Implemented a lightweight LSTM sequence model that learns temporal event patterns from historical data.
The model predicts the next likely care need, enabling proactive support and early intervention.

Challenges we ran into

Unreliable UDP transmission: We experienced packet drops and timing inconsistencies when multiple devices communicated simultaneously. Fine-tuning buffer sizes, timeout intervals, and retry mechanisms improved stability.
Hardware constraints: Fitting buttons, LEDs, and wireless components into a small, ergonomic form factor required careful pin mapping and power optimization.
Real-time synchronization: Maintaining consistent timestamps between IoT devices, the backend, and the web dashboard required implementing clock drift corrections and heartbeat pings.
Data interpretation: Translating raw events into meaningful care insights demanded thoughtful data modeling and threshold tuning for ML predictions.

Accomplishments that we’re proud of

Achieved end-to-end connectivity — from button press on an ESP32 device to instant visualization on the caregiver dashboard.
Designed a fully functional prototype combining embedded systems, cloud infrastructure, and real-time analytics.
Integrated an ML-powered predictive model that identifies recurring patterns, enhancing proactive caregiving.
Created a visually clear and emotionally intuitive UI, ensuring caregivers can respond quickly under pressure.

What we learned

Gained deep understanding of IoT communication protocols, particularly UDP and SSE.
Strengthened practical skills in backend architecture and data synchronization between devices and cloud systems.
Developed insights into lightweight machine learning deployment for small datasets and edge scenarios.
Most importantly, learned how human-centered design can guide the technical process — making technology serve emotional and social needs, not just functional ones.

What’s next

User testing: Deploy prototypes in elderly homes and caregiving centers to gather real-world feedback.
Hardware refinement: Transition from a breadboard prototype to a compact, wearable form factor resembling a wristband or badge.
Model enhancement: Improve the ML prediction accuracy through larger datasets and personalized calibration per user.
Integration with healthcare systems: Sync with medical record platforms or caregiver management systems for a holistic care workflow.
Scalability: Extend the CareSync platform to support multi-user monitoring and cross-facility analytics dashboards.

Our vision is to make caregiving environments more compassionate, data-driven, and responsive — where every signal of need, however subtle, is recognized and addressed in real time.