SHIELDHER: An AI-Powered Women Safety System with Offline GSM Emergency Alert Mechanism

Abstract

A woman who suddenly finds herself in danger rarely has the time, the calm or the opportunity to unlock a phone, open an application and decide what to do next. ShieldHer is proposed to address this challenge. When danger arises, the user presses a single button on a wearable, ESP32-based device if the hardware is available, or the SOS button of a Flutter mobile application if the phone is available. Either action starts the same emergency workflow: the location is captured and shared with the saved emergency contacts, a camera frame is taken as evidence and analysed by a YOLOv8 detector for people and weapon-class objects, and the alert is stored with its source, coordinates and status in a Flask and SQLite backend until it is resolved. A five-second cancel window limits accidental alerts, an event-identifier check removes duplicates, and a SIM800L GSM module is included in the hardware design to send SMS when no internet connection is available. ShieldHer therefore brings the wearable trigger, the mobile application and the evidence-backed alert history together as one complete and user-friendly safety system. The backend passed 28 of 28 functional tests, created an alert in a median of 2.49 ms on the server.

Introduction

ShieldHer – Smart Women’s Safety and Emergency Alert System is an IoT-based safety application designed to help users send emergency alerts quickly during dangerous situations. Traditional safety applications require users to unlock their phones, open an application, and press an SOS button, which may be difficult during emergencies. ShieldHer addresses this problem by providing two emergency activation methods: a mobile application and a physical SOS button connected to an ESP32 microcontroller.

The mobile application is developed using Flutter, while the backend uses Python Flask and an SQLite database. When the user activates the SOS button, a five-second countdown begins, allowing accidental alerts to be cancelled. Once confirmed, the system records the emergency with the user's location, timestamp, and trigger source. The physical button sends an event through Firebase Realtime Database, which is received by the mobile application and forwarded to the same backend.

The system also integrates GPS, GSM, and AI-based visual analysis. The NEO-6M GPS module provides the user's location, while the SIM800L GSM module is designed to send emergency SMS messages to saved contacts when internet connectivity is unavailable. A camera captures an image during alert generation, and the YOLOv8 model analyzes it to identify people and potential weapon-related objects.

ShieldHer maintains an emergency alert history with two statuses: ACTIVE and RESOLVED. Users can review their alerts, manage emergency contacts, share their location, and access a direct Call 112 option. Duplicate event detection helps prevent repeated alerts, while password hashing improves account security.

The project combines IoT, artificial intelligence, GPS, GSM, cloud communication, and mobile application development to create an integrated personal safety solution. The software alert workflow has been tested, while the hardware and GSM components have been designed and connected but require further testing to measure actual message delivery performance.

Conclusion

ShieldHer joins a wearable hardware button, a phone button, a location, a camera frame and an alert history into a single pipeline. The software path was tested thoroughly for a prototype: all 28 functional tests passed, alert creation took about 2.5 ms on the server, and no alert was lost in a 200-request concurrent test. Screens from the application show that a phone alert and a weapon detection both arrive as complete records with evidence attached. The hardware, GPS and GSM parts are the next thing to prove, and the detector needs a proper accuracy study. The next steps are to measure end-to-end latency from the button press to the SMS on a contact’s phone with the internet disabled, to test GPS fix time indoors and outdoors, to replace the person-count rule with a model trained on activity or distress cues using consented data, and to move the backend to authenticated, encrypted communication. Direct hand-over to emergency services would also need an official interface and should be added only with legal advice.

References

[1] D. Chand, S. Nayak, K. S. Bhat, S. Parikh, Y. Singh, and A. A. Kamath, “A mobile application for women’s safety: WoSApp,” in Proc. IEEE TENCON, 2015, pp. 1–5. [2] R. S. Yarrabothu and B. Thota, “Abhaya: An Android app for the safety of women,” in Proc. IEEE INDICON, 2015, pp. 1–4. [3] D. S. Prashanth, G. Patel, and B. Bharathi, “Research and development of a mobile based women safety application with real-time database and data-stream network,” in Proc. IEEE ICCPCT, 2017. [4] N. Saranya, R. Aakash, K. Aakash, and K. Marimuthu, “A smart friendly IoT device for women safety with GSM and GPS location tracking,” in Proc. IEEE ICECA, 2021, pp. 409–414. [5] W. Akram, M. Jain, et al., “Design of a smart safety device for women using IoT,” Procedia Comput. Sci., vol. 165, pp. 656–662, 2019. [6] A. Kane, T. Maktum, V. Mane, and N. Pulgam, “IoT based automatic women’s safety device for enhanced personal security,” in Proc. IC3Com, 2024, pp. 46–51. [7] K. Jayaram, G. Krishna K., C. Hamsani, and K. Karishma, “Modern safety device for women’s protection using IoT,” in Proc. ICRDICCT, 2025, pp. 72–77. [8] A. Joyce, S. Mohanalakshmi, P. Rohini, and S. Mohana, “SafeShe: IoT-enabled women security and well-being,” in Proc. ICRDICCT, 2025, pp. 155–160. [9] M. S. Farooq, A. Masooma, U. Omer, R. Tehseen, et al., “The role of IoT in women’s safety: A systematic literature review,” IEEE Access, 2023, doi: 10.1109/ACCESS.2023.3252903. [10] V. S. S. Bhuvaneswari and M. Thangamuthu, “Towards intelligent safety: A systematic review on assault detection and technologies,” Sensors, vol. 25, no. 13, Art. 3985, 2025. [11] P. M. Yadav, V. A. Shingare, and N. C. Yadav, “AI-based intelligent women safety and surveillance system,” TIJER, vol. 13, no. 8, pp. a622–a625, Aug. 2026. [12] G. Jocher, A. Chaurasia, and J. Qiu, “Ultralytics YOLOv8,” 2023. [Online]. Available: https://github.com/ultralytics/ultralytics [13] Google, “Flutter documentation.” [Online]. Available: https://docs.flutter.dev [14] Pallets Projects, “Flask documentation.” [Online]. Available: https://flask.palletsprojects.com [15] SQLite Consortium, “About SQLite.” [Online]. Available: https://sqlite.org/about.html [16] Espressif Systems, “ESP32 series datasheet.” [Online]. Available: https://www.espressif.com [17] OWASP Foundation, “Password storage cheat sheet,” OWASP Cheat Sheet Series. [Online]. Available: https://cheatsheetseries.owasp.org [18] R. T. Fielding, “Architectural styles and the design of network-based software architectures,” Ph.D. dissertation, Univ. California, Irvine, CA, USA, 2000.

Copyright

Copyright © 2026 Fathima G., Yackshita R., Soumya R., Swathi S.. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.