People detection is widely used in remote health monitoring and smart homes. Passive infrared (PIR) sensors are the most used devices-free detection system for their affordability, non-invasiveness, low power consumption and high accuracy. However, there are still many challenges related to analyzing digital output data to extract useful information with multi-PIR systems. Thus, in this paper we propose a measurement system composed of multi-PIR sensors mounted on the head of a mobile social robot in which the granularity of the overlapped field of view (FoV) of sensors are considered to increase the localization accuracy. In addition, a Decision Tree (DT) classifier algorithm is trained on the system data to improve the localization accuracy. The results show that the accuracy of the system is 96% for tests performed in controlled environments when subjects have gait movement constraints. The accuracy of the system decreases (83.3%) when no constraints are applied.
Localization of Older People in an Indoor Scenario: A Measurement System Based on PIR Sensors Installed in a Social Robot / Ciuffreda, I.; Casaccia, S.; Revel, G. M.. - (2023), pp. 208-212. [10.1109/MetroLivEnv56897.2023.10164045]
Localization of Older People in an Indoor Scenario: A Measurement System Based on PIR Sensors Installed in a Social Robot
Ciuffreda I.;Casaccia S.;Revel G. M.
2023-01-01
Abstract
People detection is widely used in remote health monitoring and smart homes. Passive infrared (PIR) sensors are the most used devices-free detection system for their affordability, non-invasiveness, low power consumption and high accuracy. However, there are still many challenges related to analyzing digital output data to extract useful information with multi-PIR systems. Thus, in this paper we propose a measurement system composed of multi-PIR sensors mounted on the head of a mobile social robot in which the granularity of the overlapped field of view (FoV) of sensors are considered to increase the localization accuracy. In addition, a Decision Tree (DT) classifier algorithm is trained on the system data to improve the localization accuracy. The results show that the accuracy of the system is 96% for tests performed in controlled environments when subjects have gait movement constraints. The accuracy of the system decreases (83.3%) when no constraints are applied.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
