Emergency Aid+

Inspiration

Falls are among the leading causes of injury for older adults and mobility-impaired individuals. While fall detection apps exist, most stop at sending an SOS. We wanted to go beyond that combining AI-driven fall detection, injury image recognition, and physician-backed first aid guidance into one accessible platform that empowers patients and caregivers.

What it does

Emergency Aid+ is an AI powered mobile application that:

Detects falls in real time using smartphone sensors and wearables.

Analyzes injury images with supervised learning to classify bruises, cuts, or fractures.

Provides instant first aid guidance through a Retrieval-Augmented Generation (RAG) system connected to physician-curated advice.

Escalates critical cases by auto-triggering 911 calls and SOS alerts with location sharing.

Predicts future fall risks based on gait and medical history to enable preventive care.

How we built it

Fall Detection: Trained supervised ML models on accelerometer and gyroscope data to distinguish between normal movements and actual falls.

Injury Detection: Collected and labeled injury image datasets to classify severity levels.

RAG Pipeline: Built a retrieval system connected to physician-approved first aid manuals, enabling context-specific recommendations.

Emergency Escalation: Integrated phone dialer and SOS modules for immediate response.

Challenges we ran into

Collecting high-quality, diverse injury images for supervised learning while respecting privacy and ethics.

Avoiding false positives in fall detection (e.g., sitting quickly vs. actual falls).

Designing a UI simple enough for elderly users while still offering advanced functionality.

Integrating RAG pipelines efficiently on resource-limited mobile devices.

Accomplishments that we're proud of

Creating a multi-layered safety net: prevention, detection, aftercare, and emergency escalation in one solution.

Establishing an app design that combines AI innovation with healthcare practicality.

Receiving encouraging feedback from healthcare professionals about real-world applicability.

What we learned

How to combine supervised learning and RAG techniques in a healthcare context.

The importance of human-centered design for elderly and vulnerable populations.

The need for cross-disciplinary collaboration between AI engineers, healthcare providers, and UX designers.

Balancing technical feasibility and ethical responsibility in healthcare AI.

What's next for Emergency Aid+

Expand the injury detection dataset to cover more conditions (burns, sprains, head trauma).

Partner with hospitals and insurers to integrate the app into preventive care programs.

Add multilingual voice assistance to serve global users.

Develop an offline mode for rural and low-connectivity areas.

Pilot deployment in nursing homes and community health centers

Built With

• amazon-web-services
• bedrock
• ec2
• python
• rag
• s3