Multi-sensor Platform for Detection of Anomalies in Human Sleep Patterns

This work describes a multi-sensor platform for anomalies detection in human sleep patterns. The inputs of the platform are sequences of human postures, extensively used for analysis of activities of daily living and, more in general, for human behaviour understanding. The postures are acquired by using both ambient and wearable sensors that are time-of-flight 3D vision sensor, ultra-wideband radar sensor, and three-axial accelerometer. The suggested platform aims to provide an abstraction layer with respect to the underlying sensing technologies, exploiting the postural information in common to all involved sensors (i.e., Standing, Bending, Sitting, Lying down). Furthermore, in order to fill the lack of datasets containing long-term postural sequences, which are required in human sleep analysis, a simulator of activities of daily living/postures has been proposed. The capability of the platform in providing a sensing invariant interface (i.e., abstracted from any specific sensing technology) was demonstrated by preliminary results, exhibiting high accuracy in sleep anomalies detection using the three aforementioned sensors.