| dc.description.abstracteng | The exact characterisation of planets, their host stars and the structure of their systems is an essential part of exoplanet research. This helps to understand the formation and evolution of planetary systems. Planets detected with the transiting method that show transit timing variations (TTVs) are particularly suitable for a detailed characterisation of their system. TTVs result from dynamical interactions between the system objects. Hence, from TTVs the orbital configuration and the planetary masses are determinable. Together with the radius defined by the transits, the planetary density is calculable, which helps to understand the planetary nature. This thesis is dedicated to follow-up observations and the dynamical modelling of TTV planetary systems in order to enable a refined system characterisation. Two transiting planetary systems discovered with the Kepler telescope and containing TTV planets are targets of this characterisation. To extend the observation baseline with the aim of capturing the full dynamic cycle of the TTV curves, ground-based follow-up observations of the planets transits were performed and processed within the framework of the KOINet (Kepler Object of Interest Network). To enable a comprehensive and self-consistent analysis of the systems a photodynamical model was developed for the entire photometric light curve. The photodynamical model performs a numerical integration of the entire system over the time span of observations taking into account the dynamical interactions between all objects and calculates transit light curves from the output. Kepler-9 is the first of the two systems which was subject to KOINet follow-up observations and the photodynamical analysis. The two planets b and c show anti-correlated, sinusoidal-like TTV curves. The photodynamical analysis of the system results in precise planet parameter determinations of the order of ~ 1 %, which makes them the planets with the best determined densities in the Neptune mass regime. In addition, the analysis predicts that the transits of Kepler-9c will disappear by 2050 due to orbital precession as a result of the strong interaction between the planets. Correspondingly, planet b will migrate towards the lower latitudes of the star. In the next 30 years the latitudes of the star will be scanned by the transits of the planets, where planet b will move towards possible spot regions and planet c will explore the limb of the star before disappearing. The second analysed system is Kepler-82 with the TTV-showing planets b and c. Here, the TTVs are not anti-correlated and the curve of planet c exhibits jumps every three consecutive transits, this feature is called chopping signal. The chopping signal is not induced by planet b but originated by a third outer component. With only Kepler data, two possibly system configurations are found, where an outer planet is near a 3:2 or 3:1 period ratio to planet c. The adding of KOINet follow-up observations leads to a unique solution resulting in the detection of a new non-transiting planet in the system, Kepler-82f, orbiting the star near a 3:2 commensurability to planet c. Both systems are examples of how planets in transiting systems can be missed in the light curves, since the dynamical interaction between planets can cause small deviations from co-planarity. Kepler-9c would have been missed if the Kepler mission would had been launched 40 years later and the Kepler-82 system could have shown a completely different combination of transiting planets if it had been observed at another time. The follow-up observations of the systems and their analysis with the self-consistent photodynamical model developed here enabled the precise parameter determination and system characterisation, which led to the prediction of the disappearance of the transits of Kepler-9c and to the discovery of the planet Kepler-82f. | de |