Potential Active Shooter Detection Based on Radar Micro-Doppler and Range-Doppler Analysis Using Artificial Neural Network

This paper presents a detection method of remotely identifying a potential active shooter with a concealed rifle/shotgun based on radar micro-Doppler and range-Doppler signature analysis. By studying and comparing the micro-Doppler and range-Doppler information of human subjects carrying a concealed rifle versus other similar activities, special features are extracted and applied for detecting people with suspicious behaviors. An artificial neural network is adopted in this work to complete the activity classification, and the classification result shows a 99.21% accuracy of differentiating human subjects carrying a concealed rifle from other similar activities. Due to the properties of radar sensor, the proposed method does not involve sensitive information such as visual images, and thus can better protect the privacy while being able to see-through the clothing for reliable detection.