Virtueller Zug als mobile Entwicklungs- und Validierungsumgebung für den Galileo Online Navigationsempfänger [thesis]

Andreas Trzuskowsky, Dirk Abel, Bernhard Rumpe
2021
When developing new applications, for example in the automotive industry, there are two common approaches. On the one hand, simulations can be carried out. Depending on the quality of the simulation, the expected result can be well estimated. On the other hand, there is the full-scale experimentation with a prototype. Hereby the best results can be achieved. However, prototyping is time-consuming and costly. Therefore, it is shown in this thesis, that also scaled experiments can be used to
more » ... n useful results quickly. An example for this is the development of the satellite navigation receiver for railroad applications within the project Galileo Online. The development is carried out as far as possible with model-scale test vehicles. The used test vehicles were developed at the Institute of Automatic Control at RWTH Aachen University. They are called Buggys and are about 1 m long, 90 cm wide and 30 cm high and suitable to be automated by existing real-time hardware. The Buggys have to behave like a train, so that the Galileo Online navigation receiver cannot detect a difference between a train and the Buggys. The vehicles are only allowed to move along predefined paths in order to simulate railway tracks (virtual track). In addition, several vehicles combined to form a "train" should behave as if they were mechanically coupled to each other. For this purpose, based on position measurements and inter Buggy communication, a constant distance in between the vehicles shall be maintained at all velocities (virtual coupling). The combination of virtual track and virtual coupling is called virtual train: It is developed and presented in this thesis. Two methods are investigated in order to realize the virtual train: The so-called path control and a flatness-based control. The path control, which operates at a rate of 10 Hz, is robust but computationally intensive. The flatness-based control is comparatively easy to calculate but makes higher demands on the setpoints. Therefore, a trajectory generator must be included. [...]
doi:10.18154/rwth-2022-02676 fatcat:yo2c2s4ezjba5j2vgnxbdlb4ia