ResQMesh AI Platform is an innovative open-source Android application designed to enhance emergency communication by leveraging on-device artificial intelligence (AI) within offline Bluetooth Low Energy (BLE) mesh networks. This platform is specifically tailored for disaster response, addressing scenarios where traditional communication infrastructure may be compromised due to natural disasters such as hurricanes, earthquakes, or wildfires. By enabling reliable and efficient communication without relying on the internet, ResQMesh plays a crucial role in saving lives during critical moments.

Key Features

• Decentralized BLE Mesh Networking: Operates without the need for internet connectivity or centralized servers, ensuring resilience and reliability during emergencies.
• On-Device AI Inference: Utilizes AI to make real-time decisions locally, enhancing situational awareness and response efficiency.
• FEMA-Compatible Reporting: Generates reports formatted according to U.S. federal emergency management standards, facilitating effective communication among emergency responders.

The Challenge

In disaster zones, communication breakdowns often lead to preventable casualties. Existing messaging systems typically treat all messages equally, creating system overloads and delays. The ResQMesh AI Platform addresses these issues through a multifaceted approach:

Architectural Overview

The ResQMesh AI Platform is structured around a multi-module architecture, which includes:

• AI Message Priority Classifier: A dual-stage message classifier that allocates priority to incoming messages, ensuring that critical alerts reach responders immediately.
• Offline Speech Recognition (STT): Allows users, particularly those with disabilities, to send messages using their voice without any internet requirement.
• BLE Priority Queue: This feature ensures that vital messages are transmitted before less urgent communications at the radio layer, significantly speeding up delivery times for CRITICAL messages.
• FEMA ICS-213 Situation Report Generator: Automates the creation of structured reports to keep emergency management aligned with federal guidelines.
• AI Energy Optimizer: Manages transmission power intelligently to prolong device and network operation during extended disaster conditions.

Technical Specifications

The platform is built using:

• Language: Kotlin
• UI Framework: Jetpack Compose with Material Design 3
• AI/ML Runtime: TensorFlow Lite for on-device machine learning
• Speech Recognition: Powered by Vosk, enabling offline speech-to-text capabilities
• Communication Protocol: Bluetooth Low Energy for secure and efficient message transmission

Future Development

Future enhancements will focus on expanding situational awareness through real-time mapping, federated learning for AI model improvement, and broader support for multilingual users affected by disasters.

Research & Impact

The ResQMesh AI Platform aligns with U.S. federal initiatives including the Critical and Emerging Technologies List and FEMA's strategic plan for disaster resilience, contributing to national priorities in emergency preparedness. By improving communication strategies in disaster response, this project directly addresses critical needs identified in federal research programs.

In essence, the ResQMesh AI Platform represents a significant step towards enabling effective communication in the face of adversity, enhancing the ability of communities and responders to react swiftly and efficiently during emergencies.