Velazquez, Fermin (2022)
Development of a benchmarking methodology for motion control algorithms.
Technische Universität Darmstadt
doi: 10.26083/tuprints-00021809
Bachelorarbeit, Erstveröffentlichung, Verlagsversion

Kurzbeschreibung (Abstract)

The UNICARagil project is funded by the Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung, BMBF). Along with other german universities and companies, the Institute of Automotive Engineering (Fahrzeugtechnik Darmstadt, FZD) at the Technische Universität Darmstadt is collaborating in this project whose objective is the development of fully driverless electric vehicles. The concept of the vehicle consists in a platform and add-on modules. This enables a modular approach to the development, improving the flexibility and updatability of the different components. The drivetrain structure of the project’s vehicle consists of a four-wheel independently steered (4WIS) and four-wheel independently driven (4WID), that lies within the overactuated (OA) vehicles set. This results in independent control of yaw, contrary to the traditional twowheel steered drivetrains yielding an underactuated system. The motion control system is currently being developed at the FZD. To evaluate the performance of the motion control algorithms, a benchmarking methodology is developed in this thesis. In the first place, a study of the UNICARagil project and investigation on the details of the project’s control system is performed. Then, a revision of the state of the art of the evaluation of motion control algorithms (MCAs) is conducted. A set of requirements for MCAs based on the research made is proposed. A literature review of the different metrics and test methods is performed, and their applicability within the project is investigated. Finally, an overall benchmarking methodology that allows the evaluation of MCAs is developed based on the previously defined requirements. The methodology is then applied to an existing controller within the project. To perform the evaluation of the different metrics and test methods within the benchmarking methodology, several models for the vehicle dynamics as well as for the different components were employed. The multiple-input multiple-output (MIMO) control system was decoupled into three singleinput single-output (SISO) systems after performing Relative Gain Array (RGA) to the vehicle dynamics model. For the analytical evaluation of stability, each of the components along the control system was linearized. The analysis of the accuracy was performed with a realistic model of the vehicle. The rest of the experimental tests were simulated by realistic models developed within the UNICARagil project and using IPG CarMaker

Frage zum Eintrag

Redaktionelle Details anzeigen