Work

Machine Learning Techniques for Device-Free Indoor Person Tracking

Year: 2021
Type: dissertation
Abstract: Low-cost, low-maintenance, and accurate indoor localization of persons is important for several applications, such as health care, safety monitoring, and resource usage optimization.For example, assisted living applications can lower assistance costs and improve quality of life, which are increasingly important as the projected ratio between working-age and elderly people keeps on increasing.Several smart home application can also benefit from indoor localization.For instance, room occupancy information, alongside other factors, can help reducing energy consumption by controlling the ambient temperature, lighting, and water consumption.Health care applications, such as those monitoring the human activity for extended periods, can detect behavioral changes (e.g., gait changes), which can help recognizing the early onset of diseases, e.g., Parkinson's.Moreover, presence monitoring systems can also detect unauthorized intrusions, e.g., through the house windows, which can be a sign of burglary attempt.Many localization techniques have been researched over the years.However, no silver bullet solution emerged mainly because of the complexity of person localization in indoor environments.A large variety of solutions, such as ultra-wide band or Wi-Fi techniques, assume that the person carries a device to be visible to the localization system.However, always wearing or carrying a device is not realistic for all indoor conditions and activities, reducing the acceptance of such systems.Other solutions are privacy-invasive, such as image-based systems, or may require significant infrastructure changes, such as changing floors.Systems based on infrared (IR) radiation, e.g., using passive infrared detectors (PIRs) or thermopiles, can be significantly affected by various environmental heat sources.Capacitive sensors provide unique advantages, such as simple installation, privacy, and low cost.The capacitive sensors can be used in transmit mode, shunt mode, or load mode.The transmit and shunt modes use two electrodes: the oscillator electrode and the ground electrode.The capacitive sensors operating in load mode use a single oscillator electrode, while the human body and the environment act as ground.While two-plate capacitive sensors operating in transmit and shunt modes can be more accurate, single-plate sensors working in load mode are more attractive because of their simpler installation.Capacitive sensors in load mode have been used to localize and identify persons PublicationsThe material presented in this thesis has been published in form of the following journals and conferences • Bin Tariq, O.,
Source: PORTO@Iris
Language: English
Open Access status: green
OpenAlex ID: W3194028367
Cites: 151
Cited by:
Related to: 10