AI POWERED HEATWAVE MONITORING AND SMART CROP PROTECTION SYSTEM
Because distance should never mean the absence of care...
- Problem, approach and result explained
- Expert reviewed· 6 reviews
Overview
The problem
Elderly people, especially those living alone or without constant supervision, are more vulnerable to accidental falls, sudden health emergencies and unsafe room conditions. This matters because a delay in identifying such situations can put their safety and well-being at risk. Timely monitoring can help caregivers become aware of emergencies and provide assistance when needed. We aim to develop a smart elderly monitoring and assistance bot that follows the elderly person using a wearable wristband, monitors pulse rate, detects falls through a webcam using YOLOv8n-Pose and BiLSTM, and monitors room temperature to generate LED and buzzer alerts. The goal is to provide an affordable solution that improves elderly safety and supports caregivers in timely assistance.
The approach
Approach: Components, Wiring and Challenges Our project consists of two main units: a wearable wristband and a movable bot. We selected affordable, easily available components to make the system cost-effective.
- Components Used and Why
A. Wearable Wristband
Pulse-Rate Sensor: Measures the elderly person's pulse rate. Arduino: Processes the sensor readings. nRF Transmitter: Sends data wirelessly to the bot. B. Movable Bot Arduino: Main controller of the bot. nRF Receiver: Receives wristband data and supports wireless communication. DHT11: Measures room temperature. LED: Blinks at 18°C and 16°C. Buzzer: Activates at 10°C. Webcam: Captures movements for fall detection. Motor Driver and BO Motors: Control the bot's movement. Battery: Supplies power to the bot. 2. How It Is Wired The wristband's pulse-rate sensor and nRF transmitter are connected to its Arduino. The bot's nRF receiver communicates with its Arduino through SPI connections. DHT11, LED and buzzer are connected to the bot's Arduino. The motor driver connects the Arduino to the BO motors. The webcam is connected to the processing system running the AI models. 3. Software Approach YOLOv8n-Pose: Extracts human body keypoints. BiLSTM: Analyses movement sequences to recognise walking, lying down and falling. 4. What Challenged Us the Most? Our biggest challenge was building the movable bot. Designing its structure, connecting the Arduino, motor driver and BO motors, achieving smooth movement and integrating wireless communication required considerable effort. 5. Cost-Effectiveness
We prioritised low-cost, readily available components to develop an affordable, practical and upgradeable elderly-assistance system.
The result
Does It Work in the Real World? Our project is designed to work as a practical elderly-assistance system using affordable hardware and AI-based monitoring.
Specifications Achieved: Pulse-rate monitoring through a single wearable wristband. Wireless communication between the wristband and movable bot. Room-temperature monitoring using DHT11. LED alerts at 18°C and 16°C, and buzzer alert at 10°C. Activity and fall detection using
YOLOv8n-Pose and BiLSTM. Movable bot development using a motor driver and BO motors. How We Tested It: Tested pulse-rate sensor readings and wireless communication. Checked DHT11 readings and LED/buzzer responses. Evaluated the AI models using walking, lying and falling datasets. Tested webcam-based detection and chatbot functionality. Checked bot movement and hardware integration. Conclusion: Our initial testing demonstrates the feasibility of the individual modules. Further integrated testing in real-world environments is needed to establish the system's overall reliability
Reflection
What Would We Do Differently? Improve the bot's stability, movement accuracy, wireless communication and AI model performance. 2. What's Still Unfinished? Complete wristband and bot integration. Improve the person-following mechanism. Complete real-time fall detection and emergency alerts. Integrate the admin chatbot. 3. What Did We Learn? We learned hardware integration, Arduino programming, wireless communication, machine learning and robotic development. We also understood the challenges of combining multiple technologies into one reliable system.
Submitted to Heat Intelligence at the Edge
- Arduino board used?
Yes
- Edge AI used?
Yes
- AI/ML model used
YES
- Sensors used
DHT SENSOR DS18B20 SOIL MOISTURE IR SENSOR
- Camera used
WEBCAM
- Modulino used
NONE
- Problem being solved
YES
- Arduino UNO Q features used
YES
- Report (PDF)
- Report (PDF)PDF
- What's working now?
PROPOSED WORKING APPROACH OF THE SYSTEM 1. Wearable Wristband and Pulse-Rate Monitoring 2. The elderly person will wear a wristband containing a pulse-rate sensor and nRF transmitter. 3. Pulse-rate data will be wirelessly transmitted to the nRF receiver connected to the Arduino on the mobile bot. 4. The wristband will also act as a wireless reference for the bot to locate and follow the person. 2. Temperature Monitoring and Alert System 1. A DHT11 sensor mounted on the bot will continuously monitor room temperature. 2. The Arduino will compare readings with predefined thresholds: - 18°C: LED starts blinking. - 16°C: LED alert continues. - 10°C: Buzzer gets activated (critical warning). 3. Alerts will notify nearby family members or caregivers. No heater or fan control is included. 1. Human Activity and Fall Detection 2. A webcam mounted on the bot will continuously capture the person's movements. 3. YOLOv8n-Pose will extract body keypoints, while BiLSTM will analyse consecutive movement sequences. 4. The system will recognise three activities: - Walking - Lying down/Sleeping - Falling 5. A confirmed fall will trigger an emergency alert. 6. Mobile Bot and Human Following 7. The wristband's nRF transmitter will communicate with the bot's nRF receiver. 8. The Arduino will use the wireless signal to track and follow the elderly person. 9. The webcam will independently monitor posture and movements for fall detection. 10. Admin-Based Chatbot and Emergency Alerts (Future Scope) An admin-based chatbot will provide access to: 1. Room temperature readings. 2. Pulse-rate information. 3. Activity and fall-detection status. 4. Temperature warnings. 5. Emergency notifications for caregivers or family members. 11. Overall Working Concept The wristband monitors pulse rate and enables wireless tracking. The DHT11 monitors room temperature and triggers LED/buzzer alerts. Meanwhile, the webcam and YOLOv8n-Pose–BiLSTM model mon
AI tools used
for research purpose