Back to search

National Technical University of Athens (NTUA)

Μελέτη και ανάπτυξη ευφυών αλγορίθμων και συστημάτων για την εφαρμογή της διάχυτης τηλεπισκόπησης σε υποβοηθούμενα περιβάλλοντα διαβίωσης

Abstract

dc:description

In a ubiquitous sensing environment, the approaches and the vision vary significantly according the application. However, there are three main approaches that are emerging in the literature: “Things” oriented, “Internet” oriented, Semantic oriented visions. The processing and the correlation of the available information in a ubiquitous sensing is a big challenge for the scientific community. Although the content of every application in ubiquitous sensing, especially in electronic health care, can vary one aspect remains the same the number of sources. For example, a system that collects data from a predefined number of physical and virtual sensors, is able to manage and analyze data from all the sensors. In contrast to the ubiquitous sensing technology, which envisions an environment consisting of a large number of sensors, the collection and process of all data is impossible. That’s why, the efficient correlation and evaluation of the available information is considered as big challenge. The progress in scientific areas such as telecommunications and machine learning can contribute to the implementation of the ubiquitous sensing in health care and more particularly in Ambient Assisted Living (AAL) structures. Based on the above, the current thesis studies, proposes and develops intelligent algorithms and systems in order to enable the implementation of ubiquitous sensing in e-health applications and more particularly in AAL. In Chapter 2 we studied the ubiquitous sensing technology based on the «Things» oriented vision. In this context, we proposed a network topology combining mature technologies such as Radio-frequency identification (RFID) and IPv6. This topology exploits characteristics from both technologies in terms of networking. Specifically, it uses an intermediate node for mapping to solve problems both in communication and information management level. Finally, this topology was evaluated in a general health care scenario. Furthermore, we developed an intelligent system as well as the respective algorithm for indoor human localization and navigation in AAL structure. A system that is based on RFID technology and the required technique have been developed. The system used intersecting mathematical triangles to identify (fixed and mobile) barriers and navigate the user to his/her final destination in a safe and natural way. To evaluate the solution, we developed the appropriate RFID tags and prototypes and they were tested in a number of volunteers and scenarios. Moreover, we studied the literature in Chapter 3 and we recognized the need of use a context awareness framework in ubiquitous sensing. Then, we implemented in Chapter 4 an intelligent system architecture for AAL structure. In the context of this architecture, an extension of the Semantic Sensor Network Ontology was developed, for better modeling and understanding of the available information generated by each sensor. The system was evaluated in a hybrid simulation system that allowed us to test scenarios having both physical and virtual sensors. In such an environment the system could enable the dynamic management, understanding and processing information from sensors by implementing various levels of creation and management of context information. Chapter 6 aims to study the way that fuzzy logic can be applied to data streams in order to provide a ranking a list for each sensor in an AAL environment. More specifically we build an intelligent system that implements the appropriate fuzzy sets extraction algorithms, similarity metrics and machine learning techniques. The system was evaluated in datasets from AAL structures, as these have been described in Chapter 5. Finally, a new evaluation metric was proposed reflecting better the specific problem. In the last chapter we developed an algorithm for profile generation based on the available sensors by using a context aware framework. This technique targeted to define the relationship between the needs / preferences / actions of the users and the available sensors. The goal was to exploit the available information and manage effectively the established sensor network. In order to do that, we calculated the degree of correlation of each sensor for every high level activity, based on previous performed by user low-level activities. Finally, the proposed method was evaluated in AAL data streams.

Degree

thesis:*
Grantor dc:publisher
National Technical University of Athens (NTUA)
Year dc:date
2017

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Tsirmpas, Charalampos
  • Τσίρμπας, Χαράλαμπος

Subjects

dc:subject × 20

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/41913

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Tsirmpas, Charalampos; Τσίρμπας, Χαράλαμπος. Μελέτη και ανάπτυξη ευφυών αλγορίθμων και συστημάτων για την εφαρμογή της διάχυτης τηλεπισκόπησης σε υποβοηθούμενα περιβάλλοντα διαβίωσης. National Technical University of Athens (NTUA), 2017. http://hdl.handle.net/10442/hedi/41913