| KEYWORD |
Automatic problems detection in trigger-action rules for a smart home
Parole chiave FORMAL METHODS, SMART HOME, STATECHARTS, VALIDATION
Riferimenti FULVIO CORNO
Riferimenti esterni Luigi De Russis
Gruppi di ricerca Intelligent and Interactive Systems - e-LITE
Tipo tesi RICERCA
Descrizione A considerable amount of research has been carried out towards enabling average users to customize their smart homes through trigger-action (“if... then...”) programming. This end-user programming modality creates, typically, a set of rules in the format "if an event happens, under certain conditions, then execute some actions", for example: "if it is after 5 p.m., close the shades" or "if the kitchen window is opening, and no one is at home, send a message to the house owner".
However, the execution of such rules, created by various home inhabitants in different times, can cause several problems like endless loops (e.g., one rule triggers another one, that triggers the first one, etc.), clashing rules (e.g., one rule executes an action and a following rule executes the opposite action), etc. In this way, the personalization of a smart home can be a very difficult task for an end-user, mainly due to catching and debugging such problems.
This thesis aims at creating a system for automatically detecting the possible problems present in a given set of trigger-action rules. Such rules should be represented with a proper model and translated in a suitable statechart (like the UML Statecharts, described in SCXML). Then, to verify the correct behavior of the entire set of rules, model checking technologies and methodologies can be applied. The output of this verification phase will be presented in a dedicated application, with optional suggestions to overcome the identified issues. The outcome of the thesis, if satisfying, will be made freely available as an Open Source project.
Vedi anche http://elite.polito.it/index.php/thesis/offers/271-problems-detection-in-trigger-action-rules
Scadenza validita proposta 25/02/2016
PROPONI LA TUA CANDIDATURA