Emergency Alerting and Evidence Collection System

Emergency Alerting and Evidence Collection System Abstract Personal safety during physical assault, abuse, and gender-based violence remains a critical challenge. Victims are frequently unable to unlock or operate a smartphone during an attack, and legal proceedings often suffer from a lack of objective, tamper-evident evidence. This thesis presents an integrated personal safety platform designed for low-resource and low-connectivity environments. The system consists of three main components: Standalone Wearable Device: Built on a dual-core microcontroller, featuring dual-modality distress detection that combines inertial and acoustic sensor data. During an emergency, it autonomously dispatches SMS alerts and an automated voice call with live GPS coordinates over a basic cellular network, operating entirely without a smartphone. Cryptographic Evidence Collection: The device automatically records ambient audio, encrypts it locally using AES-256 with a unique initialization vector, and computes a SHA-256 cryptographic hash stored on local memory to prevent tampering. It also includes a physical belt-clip removal detection circuit. Mobile Application and Decryptor Tool: Includes a cross-platform mobile application for Bluetooth Low Energy configuration, along with a secure web and command-line decryption utility used by authorized personnel to verify and decrypt recorded evidence. Author: Yonatan Amare Institution: Department of Computer Science, Unity University Degree: Bachelor of Science in Computer Science Project Period: November 2025 - June 2026 Technologies: Embedded C++, Microcontroller Hardware, Cellular & GPS Integration, Flutter, Python, AES-256, SHA-256