Description

Project Title

HYBRID SYSTEM FOR FALL DETECTION, ROUTINE MONITORING, AND ELDERLY ASSISTANCE

General Description

Design and develop a smart edge-based monitoring system intended for installation in the homes of elderly individuals living independently. The system combines AI-powered computer vision, wearable sensors, and environmental sensors to detect emergency situations such as falls, monitor daily routines, and identify unusual behavioral patterns that may indicate a decline in health or well-being.

All data is analyzed in real time using Python-based algorithms running on an edge device. Alerts are transmitted immediately to family members or emergency services through LoRa communication or Internet-based networks. To preserve privacy, all image processing is performed locally, and no video streams are stored or transmitted to the cloud.

Project Scope

Develop a hybrid monitoring platform that integrates computer vision, embedded systems, sensor fusion, and real-time analytics.

The system should:

* Detect falls automatically without requiring the user to activate an emergency button.
* Reduce false alarms by combining computer vision and wearable sensor data.
* Monitor daily activity patterns and identify abnormal behavior.
* Detect trends such as reduced mobility, prolonged bed occupancy, or significant routine changes.
* Generate real-time alerts during emergency situations.
* Store and analyze sensor data for long-term monitoring.
* Provide a dashboard for family members to review alerts and activity statistics.
* Ensure complete privacy by processing visual data locally and transmitting only event notifications.

Project Objectives

* Life-saving and prevention: Immediate detection of falls, one of the leading causes of injury among older adults.
* Privacy preservation: All image processing is performed on the edge device, with no video data stored or uploaded to external servers.
* Passive routine monitoring: Identification of behavioral trends that may indicate health deterioration before a critical event occurs.
* Real-time emergency response: Immediate notification of caregivers or family members when abnormal events are detected.

Student Requirements

* Teamwork
* Full attendance in weekly meetings
* High motivation
* Independent learning
* Personal responsibility
* Basic knowledge of Python programming
* Willingness to work with AI, computer vision, and embedded systems
* Ability to integrate hardware and software components

Development Tools

Software Environment:

* Python 3.10 or higher
* GitHub for version control and collaboration
* OpenCV for image processing
* MediaPipe or YOLOv8-Pose for human pose estimation
* MongoDB or TimescaleDB for data storage

Hardware Platform:

* Raspberry Pi 4/5 or Jetson Nano
* Standard USB or Web Camera

Sensors:

* IMU Sensor (MPU6050): Wearable accelerometer and gyroscope capable of measuring acceleration up to ±16g for impact detection during falls.
* PIR Motion Sensors: Installed in different rooms to monitor movement and occupancy patterns.
* BME280 Temperature Sensor: Used to detect environmental conditions that may indicate hypothermia or heat-related risks, with a resolution of 0.1°C.

Communication:

* LoRa communication modules and/or Internet-based communication infrastructure for remote alert delivery.

Research Questions

Computer Vision and Sensor Fusion:

* How can human pose estimation models be combined with wearable accelerometer measurements to minimize false-positive fall detections?
* Can Python-based algorithms distinguish in real time between normal activities, such as quickly sitting on a sofa, and actual falls?

Routine Monitoring and Privacy:

* How can routine metrics and activity statistics be presented to family members without exposing sensitive visual information?
* Can changes in room-to-room movement frequency, measured through PIR sensors, predict declining health conditions, illness onset, or mood deterioration?

Development Plan

Sprint 1 (Month 1)

* Requirements analysis and system architecture design.
* Integration of motion sensors and wearable IMU sensors with Raspberry Pi.
* Development of initial Python modules for sensor data acquisition.

Sprint 2 (Month 2)

* Integration of the camera system.
* Implementation of a pose estimation model using MediaPipe or YOLOv8-Pose.
* Development of image-based fall detection algorithms using posture transitions from vertical to horizontal positions.

Sprint 3 (Month 3)

* Development of the sensor fusion algorithm.
* Database integration using MongoDB or TimescaleDB.
* Development of the family monitoring dashboard.

Sprint 4 (Month 4)

* Testing in simulated home environments.
* Optimization for real-time performance.
* Validation of fall detection accuracy and false alarm reduction.
* Preparation of all final project deliverables.

Deliverables

* System Requirements Specification (SRS) Document describing system architecture, privacy mechanisms, AI models, and false alarm prevention strategies.
* Organized GitHub Repository containing source code, documentation, and deployment instructions.
* Functional Working Prototype including camera and sensor integration capable of detecting falls and generating live alerts.
* Family Monitoring Dashboard displaying activity metrics and notifications.
* Academic Poster presenting the technological innovation and social impact of the project.
* Final Project Presentation (PPTX).
* Demonstration Video showcasing the system in operation.

Additional Notes

* The project combines Artificial Intelligence, Computer Vision, Embedded Systems, Internet of Things (IoT), Sensor Fusion, and Healthcare Technologies.
* Students will gain hands-on experience in developing real-time AI applications on edge devices.
* The project addresses a significant social challenge by improving the safety and quality of life of elderly individuals while preserving their privacy.
* Each project team may include up to 5 students.
SafeSense: AI-Powered Fall Detection and Elderly Monitoring System

AI and Machine Learning

Description

Emphasis in project execution

The project is has cooperation with the industry and combines meeting deadlines while being creative and focused on the task
