Institutes outside Greece
Enhancing intelligent building systems using behavioural patterns
Abstract
dc:descriptionIntelligent buildings technologies are called to support the well-being of their occupants and promote independent living. On one hand, the potentially intimidating to some people idea of having electronic monitoring and control devices at home needs addressing and on the other hand, the optimal and trustworthy system operation needs to be ensured. With regards to the first issue, the author developed a graphical, intuitive and user friendly simulation tool to aid the familiarisation of users and personalised customisation of systems' design and specification. In support of system's trustworthiness, the thesis mainly focuses on the profiling of user behaviour especially in view of its expected dynamism due to changes in user roles, needs and preferences. Through unobtrusive monitoring, unhindered user movements in a real environment were recorded and authentic behavioural patterns conforming to work norms, group norms and personal habits appear beneficial to the system services' enhancements at the individual level. In support to the automated and unsupervised system services' personalisation, the present thesis differentiates from existing approaches by learning and profiling user behaviour in a manner that principally considers the non-stationarity of the observations as behaviour evolves over time. Using the example of the natural reasoning processes that humans employ to reason about another's behaviour and habits, a hybrid architecture based on Kohonen neural networks and particle filters has been engineered to track behaviour, identify habits and consequently validate the present and predict the future. Its operational and behavioural aspects are demonstrated by application on real observations acquired over a period of eighteen months. The proposed solution shows to be transparent and simple to reason, meets requirements for adaptation, delayed responsiveness to behavioural changes and support for variable constraints of accuracy and confidence and finally allows for clear and direct feedback pathways for application-specific fine tunings without the intervention of experts.
Degree
thesis:*- Grantor dc:publisher
- Institutes outside Greece
- Year dc:date
- 2007
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Agiannidis, Athanasios
- Αγιαννίδης, Αθανάσιος
Subjects
dc:subject × 21- Τεχνητή νοημοσύνη
- Συστήματα διαχείρισης κτιρίων
- Μοντελοποίηση συμπεριφοράς χρήστη
- Αυτοματισμοί κτιρίων
- Νευρωνικά δίκτυα
- Artificial intelligence
- Building management systems
- User behavior modeling
- Building automation
- Neural networks
- Les réseaux de neurones
- Intelligence artificielle
- Systèmes de gestion des bâtiments
- Modélisation du comportement des utilisateurs
- Automatisation des bâtiments
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Αυτοματισμοί και Συστήματα ελέγχου
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Automation and control systems
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/52909