Hybrid Indoor Security System based on Millimetre Wave Radar, RFID, and Face Recognition

Abstract

The goals of an efficient indoor security system are availability, integrity, confidentiality, and traceability. The objective of this research article is to reduce the crime rate that happens in closed locations, such as libraries and museums, among other significant locations. This work presents an integrated multi-sensor system for real-time people’s detection, tracking, and identification in internal locations including exhibition halls and museums.  Combining data from Ultra High Frequency (UHF) RFID, millimeter-wave (mmWave) radar (TI IWR14), and enhanced camera-based computer vision (YOLOv5-Tiny) produces consistent occupancy monitoring and intruder identification even under low-light or blocked environments.  To address latency and detection inconsistency, a Kalman Filter-based fusion technique aligns data across modalities, while edge acceleration with TensorRT enables real-time vision analysis.  The system includes a MATLAB-based GUI for visual feedback and alarms.  Compared to standard monosensor systems, our approach enhances range coverage, detection speed, and environmental durability.  Experimental findings demonstrate the framework's accuracy, low false alarm rate, and appropriateness for smart surveillance applications