Browsing by Author "Pulkkinen, Ville"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Asuntojen hintojen dynamiikkaa - rakennuskustannukset, kuluttajahinnat ja korkotaso
    (2008) Pulkkinen, Ville
     School of Business | Master's thesis
  • Kiinteän pilkun laskenta sulautetussa järjestelmässä
    (2010) Pulkkinen, Ville
     School of Science | Master's thesis
    Floating-point and fixed-point calculations in an embedded system are the focus of this study; in particular, in this research, an embedded system is defined as a valve positioner (made by Metso Automation) controlled by a microcontroller. Moreover, one of the significant causes of load in a microcontroller is known to be floating-point arithmetic. Thus the aim of this study is to create a fixed-point data type which replaces floating-point arithmetic if fixed-point arithmetic can be proven to reduce the load of the microcontroller, and retain reliable accuracy. In addition to this, a fixed-point data type must be scalable and transferable to various HW/SW-environments. The theoretical part of this thesis is based on a thorough review of earlier research and, moreover, explicates the demanding application environment of a valve package. A valve package consists of an industry valve, an actuator and a valve positioner. This thesis focuses on an ARM-based microcontroller in a valve positioner. The theoretical part explicates the relevant data type's e.g. floating-point numbers, integers and fixed-point numbers and their sizes and structures. A demonstration is given in the experimental section of how a fixed-point data type is created based on the theoretical part with an oscilloscope being used to measure the speed of floating-point and fixed-point arithmetic. Some fixed-point models are explained and one of the previous ones is chosen for further tests. The criteria for a chosen fixed-point model are extensibility and accuracy. A chosen fixed-point model is verified against accuracy and speed. This thesis demonstrates the processes of floating-point and fixed-point calculation as well as their requirements. Finally, the thesis proposes some ideas for further development of fixed-point calculation.