ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

Identification of persons entering through the door from the accelerometers data

Tadej Vodopivec (2013) Identification of persons entering through the door from the accelerometers data. EngD thesis.

Download (65Mb)


    This thesis presents a system for automatic identification of a person who walks through a door based on the information gathered with acceleration sensors, which are mounted on the door. The proccedure for automatic identification consists of four steps. In the first step, data acquisition and coordinate system rotation is performed. Coordinate system rotation is needed in order to match the direction of measured accelerations with the direction of components of the forces acting on the door. The rotational matrix needs to be computed from dedicated measurement prior to the identification procedure. In the second step, angular acceleration, angular speed, standard deviation of the acceleration in the direction of gravity and the opening angle of the door are calculated from the rotated accelerations. In the third step, the borders between various phases of transition trough the door are identified. Integral, linear interpolation and extreme values are computed for each component of acceleration and duration is calculated for each phase. In the fourth step, classification is performed using values computed in the previous step. The system will be used in the project e-Doorman, which is developed at »Jožef Stefan« Institute. The aim of the project is to develop a prototype of electronic doorman: an intelligent system based on a door with electro-mechanic lock, tablet computer, Arduino micro-controller, and an array of sensors that offers services similar to a human doorman, improves security, and increases user comfort. The sensors gather data about events related to the door, which are used for context based reasoning and awareness achieved by artificial intelligence methods running on the tablet computer. The e-Doorman system is able to recognize the users, detect unusual entries and exits, break-ins and break-in attempts, predict user presence, offers personalized services (customizable notifications and alarms, relevant information, voice messages, greetings and tips), and remote control using intuitive graphical user interface or virtual assistant that understands natural language.

    Item Type: Thesis (EngD thesis)
    Keywords: identification,door,entrance,sensors,machine learning,signal processing
    Number of Pages: 72
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    doc. dr. Peter Peer294Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=10166868)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 2173
    Date Deposited: 19 Sep 2013 11:03
    Last Modified: 03 Oct 2013 14:16
    URI: http://eprints.fri.uni-lj.si/id/eprint/2173

    Actions (login required)

    View Item