A Novel Contactless Concentration Evaluation Technique Using Millimeter-Wave Radar

Since millimeter-wave radar has made significant progress in motion tracking and vital signs detection, it is poised to delve into further activity monitoring applications, such as concentration evaluation. When radar senses the surroundings, the object is dissected into a series of scattering points. The distribution of these points exhibits distinct patterns across varying levels of activity concentration. Based on these trends, a novel millimeter-wave concentration evaluation technique is proposed. Firstly, the relative movement feature is extracted from the signal. Then the Huber function is designed to compress the feature matrix into a feature vector. Finally, support vector regression is employed to evaluate a reasonable concentration score. In the experiments, high-concentration activities such as reading and typing are scored 92% and 88%, respectively. Conversely, middle and low concentration activities such as playing with toys and shaking shoulders are scored 76% and 59%, respectively, showcasing the technique's potential for evaluating concentration levels.