Gesture-Based Messaging for Assistive Care
A wearable that turns four tilt gestures into messages — food, water, help, emergency — for elderly and disabled people.
- Role
- Team member (3-person team)
- Team
- 3 students — IUB
- Timeline
- Jun 2025 – Aug 2025
- Organisation
- Independent University, Bangladesh
- Code
- Private repository
- Reading time
- 1 min
Overview
People who cannot easily speak or reach a phone still need to ask for help. A small wearable maps tilts to messages that appear, with a buzzer, on a carer’s receiver.
The problem
- Interaction must need almost no fine motor control.
- The link must be reliable across a home.
Context & constraints
An MPU6050 sensor was tried first and dropped as too noisy.
Approach
- Sender — ESP32 with an MPU9250 IMU; four tilt gestures detected by thresholds (left, right, forward, backward).
- Link — Bluetooth Classic in master/slave mode with automatic reconnection.
- Receiver — ESP32 with an OLED display and a 5 V buzzer.
- Messages — left: Need food, right: Need water, forward: Call help, backward: Emergency.
- Build — 3.7 V cells with TP4056 charging and 3D-printed enclosures.
Key capabilities
- Four-gesture vocabulary
- Bluetooth auto-reconnect
- Visual + audible alerts
- Rechargeable, 3D-printed units
Architecture
Wearable sender
Receiver
Two microcontrollers and one radio link. Select a component to trace its connections.
Challenges
Noisy motion data
Problem. The first sensor produced too much noise for reliable thresholds.
Solution. Switching to an MPU9250 made gesture detection stable without machine learning.
Outcome
- Reliable range
- 10–20 m
- Bluetooth Classic
Core features complete; custom messages, logging, web integration and SMS were planned as next steps.
Lessons
- The simplest reliable signal beats a clever noisy one.
Building something similar?
Tell me what you’re working on — I reply to every serious message.
Related work
Automated Weeding Robot
An Arduino Mega robot with ultrasonic obstacle sensing, GPS waypoint navigation, differential drive and a high-speed cutter. Removed 85% of weeds in the lab and 75% in the field.
- C / C++
- Arduino
Smart Aquaponics with a Digital Twin
A course research paper (3rd of 4 authors) designing ESP32 sensor nodes for water quality, a Firebase telemetry backbone, a React monitoring and control dashboard, a digital twin for scenario testing and LightGBM-based anomaly prediction.
- Python
- C / C++
- React
- Firebase & Firestore
- ESP32