The Compressed Baryonic Matter (CBM) Experiment will investigate heavy ion collisions and reactions at interaction rates of 100 kHz in a targeted energy range of up to 11 AGeV for systems such as gold-gold or lead-lead. It will be one of the major scientific experiments of the Facility for Antiproton and Ion Research in Europe (FAIR) currently under construction at the site of the GSI Helmholtzzentrum für Schwerionenforschung (GSI) in Darmstadt, Germany. CBM is going to be a fixed target experiment consisting of a superconducting magnet, multiple detectors of various types, and high-performance computing for online event reconstruction and selection. The detector closest to the interaction point of the experiment will be the Micro Vertex Detector (MVD). Consisting of four planar stations equipped with custom CMOS pixel sensors, it will allow to reconstruct the primary vertex with high precision and will help to reconstruct secondary vertices and identify particles originating from conversion in the detector material. Due to the high interaction rates foreseen for CBM, understanding and minimizing systematic errors due to the detectors’ operating conditions will become all the more important to obtain significant measurement results, as statistical errors in the measurements of many observables are diminishing due to the enormous amount of data available. Furthermore, the MVD will be the first detector based on CMOS pixel sensors used in a large physics experiment, that will be operated in vacuum. As a result, many aspects of the mechanical and electrical integration of the detector require careful testing and validation. This thesis addresses both those challenges specifically for the Micro Vertex Detector with the development of a control system for the operation and validation of the MVD prototype “PRESTO” in vacuum. The prototype was selected as device under test as the final MVD is not yet built. The developed control system helps a) to operate the prototype safely and keep it at the desired working point and b) to record important time-series data of the state of the detector prototype. Those two aspects allow the control system (which might later serve as a ‘blueprint’ for the final detector) to minimize the mentioned systematic errors as much as possible and to contribute to the understanding of remaining systematic errors using correlations with the time-series data. The controlled operation of the prototype in vacuum allowed to validate the integration concepts from a wide range of mechanical and electrical aspects in an endurance test for more than a year with 24/7 operation. The prototype for this study itself was named “PRESTO” (standing for ‘PREcursor of the Second sTatiOn of the CBM-MVD’). It represents one quadrant of an MVD detector plane, equipped with a total of 15 MIMOSA-26 sensors on the front and back side of a carrier plate. Within this thesis, major parts of the prototype itself were designed. Custom ultra-thin flat flexible cables for data and power were designed and validated. Furthermore, the CNC-machined Aluminium heatsink to mount and cool the prototype design was refined to increase thermal performance. A custom vacuum feedthrough for a total of 21 flat ribbon cables was designed and fabricated. The read-out chain for MIMOSIS-26 was extended to cover a total of 8 sensors with a single and newer TRB-3 FPGA board and was set-up with the prototype. Vacuum equipment including chambers, hoses, pumps, valves and gauges were integrated to form a large vacuum testing system. A cooling circuit for the prototype was assembled comprising an external chiller, hoses, vacuum feedthroughs, as well as temperature, flow and pressure sensors. The control system was developed to serve the needs of the prototype, while taking the requirements of the final MVD already into account. The main design goals of the control system are: • compatibility with the other detectors and the overall CBM experiment, • access to real-time measurements of all necessary parameters (‘process values’), • reliable, fail-safe operation of the detector, • recording of all time-series data (‘archiving’), • cost efficiency and acceptance within the physics community, • good usability for the users (‘operators’), • long-term maintainability. The recorded time-series data of the process variables (i.e. sensor readings) allow a post-measurement analysis of variations in the detector performance. The longterm archiving of all relevant system parameters is therefore of outstanding importance, which is why the software intended for this purpose – called “archiver” – was given special attention in this thesis. For this reason in particular, it is necessary to implement a comprehensive control system that allows the detector to be operated safely under these conditions and cooled effectively. Before the start of this doctoral thesis, vigilant and extensively trained operators were always necessary for this. The control system that has been developed makes it possible that, after basic training, the detector can also be operated by a less specialised shift supervisor during measurement campaigns. ...