STM32-Based Landmine Detection and GPS Localization Robotic Vehicle

Abstract

Autonomous landmine detection remains a major safety challenge, as conventional manual methods expose personnel to hazardous environments, take considerable time, and may offer limited spatial tracking of detected threats. These limitations are particularly significant in military, humanitarian demining, and hazardous-terrain operations where accurate localization, autonomous navigation, and real-time monitoring are required. This thesis presents an Autonomous GNSS-Guided Mine Detection and Geofenced Monitoring Robot designed to autonomously survey a defined area and identify and spatially record potential metallic mine locations. The system provides dedicated access for field operators and administrators, where operators monitor robot movement, detected targets, and surveyed regions while administrators manage operational data, geofencing, and monitoring functions. The core mechanism combines GNSS positioning with centimeter-level correction, geofencing, barometric sensing, metal detection, and motor control to continuously calculate robot position and constrain movement within the predefined boundary, providing transparent rule-based navigation rather than relying on a black-box navigation model. Supporting features include autonomous movement, virtual fencing, boundary violation prevention, metal-detection marking, real-time position tracking, sensor-status monitoring, location logging, historical data visualization, and remote dashboard monitoring.

Introduction

The text presents the development of an autonomous, GNSS-guided, geofenced, and cloud-connected robot for metal-object detection and hazardous-area surveying. Advances in robotics, satellite navigation, embedded systems, artificial intelligence, and cloud computing have created opportunities to reduce human exposure to dangerous environments and improve the efficiency of landmine-survey operations.

The literature review highlights several important studies. Kambhampati and Rajasekharan (2007) identified slow detection and false alarms as major limitations of conventional metal detectors. Dakulovi? and Petrovi? (2012) investigated complete-coverage path planning for demining robots, while Dasgupta et al. (2015) demonstrated the benefits of multi-robot coordination, sensor fusion, and improved localization. Joo et al. (2022) explored machine-learning-based metal detection using a drone, and Baur et al. (2021) investigated deep learning for detecting scatterable landmines from aerial imagery. Panwar et al. (2024) reviewed advances in robotic mobility, sensor integration, autonomous navigation, and machine learning. Collectively, these studies demonstrate the potential of automation while highlighting the need for reliable localization, effective sensor integration, and systematic spatial documentation.

The proposed project addresses these challenges by integrating an ESP32 microcontroller, GNSS positioning with centimeter-level correction, a metal detector, a barometric sensor, an L298N motor driver, and a battery-powered robotic platform. A rule-based navigation system continuously compares the robot's position with a predefined virtual boundary, allowing movement within the authorized survey area and stopping the robot when boundary violations are detected. This approach is selected because its decisions are transparent, predictable, and easier to test than those of a complex, black-box navigation model.

The project also incorporates Firebase for cloud-based data storage and real-time synchronization, along with a ReactJS web interface for monitoring robot movement, configuring geofences, visualizing detected metallic-object locations, and reviewing historical survey records. Authentication and role-based access control help ensure that only authorized users can access operational information and configuration functions. The primary objectives are to develop the integrated hardware and software architecture, implement autonomous navigation and geofence enforcement, record detection coordinates, establish cloud communication, and evaluate system performance under controlled conditions.

The dataset consists of operational records collected during survey experiments, including latitude, longitude, timestamps, GNSS status, metal-detector activity, detection events, geofence status, robot movement, and sensor information. These data are transmitted to Firebase and displayed through the web interface, enabling real-time monitoring and historical analysis.

The project has several limitations. The current navigation system relies on predefined rules and does not include advanced obstacle avoidance or machine-learning-based terrain interpretation. The metal detector can identify metallic objects but cannot independently distinguish landmines from harmless metallic debris. GNSS accuracy may also be affected by signal obstruction and environmental conditions, while cloud synchronization depends on network availability. Furthermore, the prototype requires broader testing across different terrains, soil conditions, and target types.

Conclusion

This project presents an autonomous GNSS-guided robotic platform for metallic-object detection, location recording, geofenced operation, and remote monitoring. It combines an ESP32 controller with GNSS, metal detection, barometric sensing, motor control, Firebase cloud services, and a ReactJS-based web interface. The proposed architecture addresses several limitations of conventional manual surveying by providing autonomous movement, geographic boundary enforcement, real-time position tracking, detection-location recording, cloud synchronization, and map-based visualization. The use of explicit rule-based navigation makes the movement decisions transparent and easier to test and validate. An important feature of the system is the separation between metal detection and mine confirmation. The robot records a detector event as a potential metallic target rather than claiming that every detected object is a landmine. This distinction is important because metallic debris and other objects can produce detector responses. The project therefore provides a foundation for safer and more systematic robotic surveying in controlled hazardous environments. Further improvements in positioning, obstacle avoidance, sensor fusion, target classification, offline synchronization, and large-scale field validation can extend the capabilities of the proposed platform.

References

[1] Kambhampati, S., and Rajasekharan, S., “Robotic Platforms and Sensor Technologies for Landmine Detection,” 2007. [2] Dakulovi?, M., and Petrovi?, I., “Complete Coverage Path Planning of Mobile Robots for Humanitarian Demining,” 2012. [3] Dasgupta, P., et al., “COMRADE: AMultirobot Autonomous Landmine Detection System,” 2015. [4] Joo, et al., “Drone-Operated Metal Detection Using Optimized PID Control and Machine Learning Classification,” 2022. [5] Baur, et al., “UAV-Based Landmine Detection Using Region-Based Convolutional Neural Networks,” 2021. [6] Panwar, V. S., et al., “Landmine Detection Robots: AReview of Mobility, Sensor Integration, Autonomous Navigation and Machine Learning,” 2024. [7] Yoon, and Lee, “Compact RTK-GNSS Localization for Outdoor Mobile Robots,” 2024. [8] Cho, et al., “Machine-Learning Classification of Metallic Landmines and Debris Using UAV Magnetometer Data,” 2026. [9] Firebase Documentation, “Firebase Realtime Database and Security Rules.” [10] Relevant ESP32/Firebase robotic-control research, “Cloud-Based Robotic Control and Communication Using ESP32 and Firebase

Copyright

Copyright © 2026 Prof. Ankita Dhenge, Manmeet Singh Dhillon, Taran Gondane, Avinash Sonwane. 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.