@phdthesis{discovery10112785, school = {UCL (University College London)}, year = {1990}, note = {Thesis digitised by ProQuest.}, title = {Pattern recognition in the ZEUS Central Tracking Detector}, keywords = {Pure sciences}, url = {https://discovery-pp.ucl.ac.uk/id/eprint/10112785/}, abstract = {The HERA accelerator will collide 30 GeV electrons with 820 GeV protons. The ZEUS and Hi experiments are currently being prepared to study the resulting interactions. At the heart of ZEUS, the Central Tracking Detector (CTD) is a large cylindrical drift chamber. Using a mixture of axial and stereo sense wires it will allow the reconstruction of charged particle tracks in three dimensions. The first step in full CTD track reconstruction will be to find tracks in just two dimensions using information from the axial sense wires only. In normal operation this two dimensional pattern recognition problem must be solved separately at two different stages of processing: a fast online system must provide information for trigger processing, while later a highly efficient offline system must allow as much information as possible to be extracted from the recorded events. The CTD online pattern recognition system is part of the ZEUS second level trigger (SLT) and is required to process one event every millisecond. Our solution to this difficult problem involves a highly parallel algorithm running on a large array of transputers. Our discussion of this question is broadened to include similar applications in the readout and trigger systems of future detectors. The CTD offline pattern recognition methodology bases track finding on the extension towards the interaction point of seed tracks found in the outer layers of the CTD. A high degree of software engineering ensures that our design and implementation is both robust and flexible. Several complementary options for finding seed tracks are supported, including one specially designed to exploit vector architectures. As one possible use of the CTD in early physics analysis a study is made of using tracking information to help separate the neutral current and charged current event classes.}, author = {Shaw, David} }