Files
Abstract
Seismic data has been historically used in seismology for geological, structural and military
monitoring applications. Due to all activities, natural and human-made, generate vibration,
these systems have been incorporated in new research areas. In this dissertation, several
security and health monitoring systems are proposed. These approaches can be applied
to both, the analysis of seismic movements on earth and structures and the analysis of
movements made by human beings for health and security. The whole philosophy is based on
real-time signal processing analysis and machine learning techniques over sensor networks.
Processing times, bandwidth, and the energy usage of the network are optimized taken into
consideration the computational capabilities of smart sensors to do in-situ processing. A
real-time and energy-efficient seismic sensing system is presented. It locates earthquakes
in outdoor scenarios based on sensor networks, which introduce an efficient and on-time
system for security purposes. Then, a system that consists in a smart network of sensors
is introduced. It can sense signals to recognize people, identify fall downs, and locate these
fall-downs through ambient structural vibration, which may directly impact security (intruder
detection) and health applications. Finally, a novel smart sensor system for sleep monitoring
is proposed. It is based only on the vibration produced by a body in a bed, the system can
infer sleep activities (entry/exit of bed, movement, and posture changes) and vital signs (heart
rate and respiration rate). The proposed systems represent novel and real-time solutions for
security and health, using only seismic signals.
monitoring applications. Due to all activities, natural and human-made, generate vibration,
these systems have been incorporated in new research areas. In this dissertation, several
security and health monitoring systems are proposed. These approaches can be applied
to both, the analysis of seismic movements on earth and structures and the analysis of
movements made by human beings for health and security. The whole philosophy is based on
real-time signal processing analysis and machine learning techniques over sensor networks.
Processing times, bandwidth, and the energy usage of the network are optimized taken into
consideration the computational capabilities of smart sensors to do in-situ processing. A
real-time and energy-efficient seismic sensing system is presented. It locates earthquakes
in outdoor scenarios based on sensor networks, which introduce an efficient and on-time
system for security purposes. Then, a system that consists in a smart network of sensors
is introduced. It can sense signals to recognize people, identify fall downs, and locate these
fall-downs through ambient structural vibration, which may directly impact security (intruder
detection) and health applications. Finally, a novel smart sensor system for sleep monitoring
is proposed. It is based only on the vibration produced by a body in a bed, the system can
infer sleep activities (entry/exit of bed, movement, and posture changes) and vital signs (heart
rate and respiration rate). The proposed systems represent novel and real-time solutions for
security and health, using only seismic signals.