Machine Learning Driven Multiple Human Activity Recognition and Tracking Using Radar Technology: A Review

In recent years, human detection and tracking have become crucial in military, search and rescue, and law enforcement agencies. Using machine learning techniques such as two-dimensional convolutional neural network (CNN), support vector machine (SVM), and ResNet in correct detection proves to increase efficiency with a great magnitude. This paper reviews the vast efficient methodologies used in data extraction and signal pre-processing for the effective detection and tracking of multiple human activities. The systematic reviews and quantitative analysis of existing research are thoroughly reviewed. Furthermore, it discusses the assessment and evaluation criteria based on the results extracted to showcase the challenges and opportunities of the methodologies discussed. Moreover, this study also presents the effective techniques for multiple human detection, tracking within a room and through the wall scenarios.