Pattern recognition in the Silicon Vertex Detector of the Belle II experiment

The purpose of this thesis is to present a possible way of track finding specialized for low momentum tracks within the Silicon Vertex Detector (SVD) of the Belle II experiment. Track finding algorithms are an essential tool for the reconstruction of tracks of elementary particles moving through the detector. Chapter 1.2 is a general description of accelerators used for particle collision experiments and detectors used for recording tracks of produced particles. Chapter 1.3 gives an overview of the steps of event reconstruction. Chapter 2 describes the current situation at Belle and the future Belle II. 2.1 includes some important physical aspects and their effect on the design of the experiment. 2.2 explains the concept of the superKEKB-Accelerator used for this experiment and shows how the design has been adapted to current purposes. 2.3 gives an overview of the detector and explains some details important to pattern recognition. Chapter 3 starts with a list of requirements for pattern recognition defined by the current situation in 3.1 and 3.2. Furthermore it presents some typical algorithms used for track finding and explains their advantages and disadvantages. In particular, conformal mapping 3.3, Hough transform 3.4, the Kalman filter 3.5 and the Hopfield network 3.6 are discussed. Chapter 3.7 contains our current concept for the low momentum track finder in full detail. Finally, its performance is compared with common techniques in 4.