Detection and classification of movements captured by radar for smart homes applications

• Autor(a) principal: Couto, Tiago André Martins
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/41940
• topic: Smart homes applications; Frequency modulated continuous wave radar; Movement detection and classification; Doppler effect; Machine learning

Smart homes are homes equipped with advanced automation and control systems, encompassing a variety of functions and devices, such as lighting, security, entertainment and temperature regulation, with the aim of offering greater comfort to their occupants. In such a wide range of applications, the detection and classification of movement is of remarkable importance, allowing, for example, the adaptation of temperature based on the type of movement made by a person. This dissertation presents a system for detection and classification movements captured by radar for Smart Home applications. For this purpose, a frequency modulated continuous wave radar operating in the 77 - 81 GHz band is used, where it was possible to obtain the person location and movement velocity where the person was detected. The movements were classified using Machine Learning algorithms such as K - Nearest Neighbours, Support Vector Machine, Linear Discriminant Analysis and Random Forest. To this end, data was collected from 10 different people in order to obtain an extensive data set. The positioning of the radar in the division was also studied in order to have good data collection. The data was processed using Matlab, where characteristics were created to differentiate the movements to be classified, for example, the Doppler values or speed that a person has in different types of movement. The performance of each Machine Learning algorithm used was studied.