A Conceptual Framework for Multi-Sensor Human Detection in Intelligent Perimeter Security Systems

Abstract

In an era increasingly defined by smart technologies and real-time data-driven systems, securing perimeters against unauthorized intrusions has become paramount. This research proposes a low-cost, intelligent, and power-efficient perimeter protection system that integrates multi-modal sensors—namely the HLK-LD2410B millimetre-wave radar and the Adafruit AMG8833 thermal infrared sensor—with image capture via the ESP32-CAM and centralized control through the ESP8266 microcontroller. The core novelty lies in a dual-stage machine learning pipeline that fuses thermal and visual data to accurately detect human presence while minimizing false positives. The system operates on a hierarchical decision framework, initiating with radar motion detection, followed by conditional thermal sensing, and concluding with CNN-based image validation. Upon confirmed detection, alerts are issued both locally via a buzzer and remotely through Wi-Fi-enabled notifications. Simulation and field tests demonstrate over 97% accuracy, sub-300 ms latency, and high energy efficiency, validating the system\'s suitability for smart surveillance, elderly monitoring, and adaptive home automation applications.

Introduction

This research presents an advanced, AI-powered perimeter protection system designed to enhance security across various settings by combining multiple sensor technologies and machine learning. Traditional systems like CCTV and PIR sensors often struggle with accuracy and adaptability, especially in complex or low-light environments. To address this, the proposed system integrates a high-precision radar sensor (HLK-LD2410B), a thermal infrared sensor (Adafruit AMG8833), an ESP32-CAM for visual confirmation, and an ESP8266 Wi-Fi module for communication.

The system uses a sensor fusion approach where the radar sensor continuously monitors for motion. When potential human presence is detected, the thermal sensor is activated to capture heat signatures, followed by visual image capture for further verification. Dual-stage machine learning models analyze thermal and visual data to accurately distinguish humans from other sources, minimizing false alarms. Upon confirmed intrusion, the system triggers an audible alarm locally and sends real-time alerts with images to users via a mobile or web app, ensuring both immediate deterrence and remote monitoring.

The methodology emphasizes energy efficiency through conditional sensor activation and detailed event logging for security audits. Extensive testing confirms the system’s robustness under varied environmental conditions. This integrated approach combining AI, IoT, and sensor fusion offers a scalable, cost-effective solution that improves perimeter security beyond traditional methods.

Conclusion

This research presented the design, development, and validation of an intelligent, multimodal presence detection system that integrates an AMG8833 thermal infrared sensor, a HLK-LD2410B mmWave radar module, an ESP32 microcontroller, an ESP32-CAM module, and a buzzer-based alert system. The aim was to create a reliable, responsive, and power-efficient solution for detecting human presence in indoor environments with applications ranging from smart surveillance to elderly care and room automation. The proposed system leverages sensor fusion, combining thermal and motion detection to overcome the individual limitations of each sensing modality. The AMG8833 provides thermal mapping with an 8×8 grid resolution, capable of identifying heat signatures associated with the human body, while the HLK-LD2410B radar module offers high-precision detection of movement, even in low visibility conditions. The ESP32 microcontroller processes the incoming data and implements the detection logic, whereas the ESP32-CAM captures images upon validated presence detection, and the buzzer serves as an immediate acoustic alert. Simulation results confirmed that the system logic performed correctly under varied scenarios, accurately identifying the presence of humans and avoiding false positives from environmental noise or non-human heat sources. Real-world testing demonstrated a detection accuracy exceeding 97%, with a false positive rate below 5%, and an average response latency under 300 milliseconds. The ESP32-CAM achieved a 95% image capture success rate, and power consumption remained within acceptable limits for continuous operation. This project successfully demonstrated that combining multiple low-cost sensors under a unified logic framework substantially enhances detection reliability and system intelligence. The modular architecture of the system also ensures scalability and adaptability to various IoT applications, including home automation, security surveillance, and health monitoring. In conclusion, the system offers a low-cost, robust, and efficient solution for intelligent human presence detection, providing strong potential for real-world deployment in smart environments. Future work may involve integrating cloud connectivity for remote monitoring, implementing machine learning models for behavioral pattern analysis, and optimizing the design for embedded low-power applications.

References

[1] Espressif Systems, \"ESP32 Technical Reference Manual,\" [Online]. Available: https://www.espressif.com [2] Panasonic, \"AMG8833 Grid-EYE Infrared Array Sensor Datasheet,\" [Online]. Available: https://na.industrial.panasonic.com [3] HLK, \"LD2410B mmWave Human Presence Sensor Datasheet,\" Hi-Link, [online]. Available: https://www.hlktech.com [4] Espressif Systems, \"ESP32-CAM Technical Specifications,\" [Online]. Available: https://docs.espressif.com [5] A. Zanella et al., “Internet of Things for Smart Cities,” IEEE IoT Journal, vol. 1, no. 1, pp. 22–32, Feb. 2014. [6] K. Ashton, “That ‘Internet of Things’ Thing,” RFID Journal, 2009. [7] F. Xia, L. T. Yang, L. Wang, and A. Vinel, “Internet of Things,” International Journal of Communication Systems, 2012. [8] M. Palattella et al., “Standardized Protocol Stack for the Internet of (Important) Things,” IEEE Communications Surveys & Tutorials, vol. 15, no. 3, pp. 1389–1406, 2013. [9] S. Ziegler et al., “IoT Platforms for Smart Cities,” Enabling Technologies, vol. 13, no. 1, 2020. [10] L. G. Jaimes, I. J. Perez, “Human Detection Techniques in Intelligent Surveillance,” Sensors, vol. 12, no. 2, pp. 1474–1521, 2012. [11] N. Ghosh et al., “Human Activity Monitoring using Thermal Imaging,” Procedia Computer Science, vol. 58, pp. 530–537, 2015. [12] R. Lu et al., “mmWave Radar-Based Human Detection,” IEEE Sensors Journal, 2020. [13] M. Khan et al., “Thermal Infrared Sensors for Occupancy Detection,” IEEE Transactions on Smart Grid, 2021. [14] A. Goldsmith, Wireless Communications, Cambridge University Press, 2005. [15] B. Krose, A. Bikker, “Occupancy Detection with Radar and Infrared Sensors,” Smart Homes Conference, 2016. [16] G. Fortino et al., “Internet of Things Based Smart Environments,” Springer IoT Book Series, 2014. [17] M. Weiser, “The Computer for the 21st Century,” Scientific American, 1991. [18] C. Doukas, I. Maglogiannis, “Bringing IoT and Wearable Devices into Elder Care,” IEEE EMBS, 2012. [19] K. Bouchard et al., “Ambient Intelligence for Health Monitoring,” Sensors, vol. 20, 2020. [20] D. Salim et al., “ESP32-based IoT Smart Surveillance,” International Conference on Smart Applications, 2022. [21] Y. Liu et al., “Sensor Fusion for Presence Detection,” IEEE Access, vol. 7, pp. 109-123, 2019. [22] A. Rahman et al., “Smart Home Energy Management using Sensor Networks,” IEEE Systems Journal, 2018. [23] V. I. Lakshmi et al., “Smart Home Automation Using ESP32 and Blynk,” IRJET, vol. 6, no. 5, 2019. [24] M. N. Islam et al., “IoT-Enabled Fall Detection for Elderly Care,” IEEE IoT Journal, 2020. [25] T. Moe et al., “AI at the Edge with ESP32,” arXiv preprint arXiv:2101.02802, 2021. [26] R. Yuce, “Implementation of Wireless Body Area Networks for Healthcare Systems,” Sensors, 2020. [27] H. Lu et al., “Edge Computing Framework for Real-Time Human Detection,” IEEE Access, 2019. [28] M. Alam et al., “Sensor-Based Ambient Assisted Living,” IEEE Communications Magazine, 2011. [29] M. Mazzei et al., “Thermal and Radar Fusion for Smart Buildings,” Energy and Buildings, 2019. [30] L. M. Ang et al., “Review of Smart Homes Technologies,” Sensors, vol. 20, 2020. [31] K. Gill et al., “IoT for Smart Home Automation,” IEEE Consumer Electronics, 2009. [32] A. Mahapatra et al., “AI-Powered Home Security using ESP32-CAM,” IJERT, vol. 9, 2020. [33] S. Mitra et al., “ESP32 in IoT Projects: Low Power Solutions,” Microelectronics Journal, 2021. [34] M. Wang et al., “Survey on Radar-Based Human Activity Recognition,” IEEE Sensors, 2020. [35] J. Redmon et al., “YOLO: Real-Time Object Detection,” CVPR, 2016. [36] A. Karpathy, “Deep Learning for Vision and Robotics,” Stanford CS231n, 2019. [37] N. Banerjee et al., “Smart Room Automation using Multimodal Sensing,” ICSESS, 2021. [38] J. Wu et al., “Smart Occupancy Detection using Machine Learning,” IEEE Transactions on Automation Science, 2018. [39] S. Alam et al., “Thermal Camera Based Security System,” IEEE WIECON, 2019. [40] A. Krizhevsky et al., “ImageNet Classification with Deep Convolutional Neural Networks,” NeurIPS, 2012.

Copyright

Copyright © 2025 Abhirup De, A. P. Amit Nigam , Sraman Chatterjee, Amrita Roy, Srotoswini Sen. 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.