The purpose of this research was to develop an algorithm for a wearable device that would prevent people from drowning in swimming pools. The device should detect pre-drowning symptoms and alert the rescue staff. The proposed detection method is based on analyzing real-time data collected from a set of sensors, including a pulse oximeter. The pulse oximetry technique is used for measuring the heart rate and oxygen saturation in the subject’s blood. It is an optical method; subsequently, the measurements obtained this way are highly sensitive to interference from the subject’s motion. To eliminate noise caused by the subject’s movement, accelerometer data were used in the system. If the acceleration sensor does not detect movement, a biosensor is activated, and an analysis of selected physiological parameters is performed. Such a setup of the algorithm allows the device to distinguish situations in which the person rests and does not move from situations in which the examined person has lost consciousness and has begun to drown.
The paper describes a signal preprocessing approach intended for a multistatic, multiband passive radar system. Each transmitter broadcasts signals in different but closely spaced spectral channels, whereas each receiver has only two physical channels covering whole spectrum span of interest. In such scenario, the reference signal for particular spectral channel can be taken from the most advantageous receiver physical channel and used offline to perform range-Doppler processing of signals captured in other channels. To allow that, processing steps such as signal channelization and multisite synchronization have been proposed. The paper includes results of airplane detection, localisation and tracking obtained for real measurement data based on DVB-T illuminators.