h1

TY  - THES
AU  - Shahirpour, Arash
TI  - Digital twin and trajectory tracking for articulated dump trucks
PB  - Rheinisch-Westfälische Technische Hochschule Aachen
VL  - Dissertation
CY  - Aachen
M1  - RWTH-2025-09517
SP  - 1 Online-Ressource : Illustrationen
PY  - 2025
N1  - Veröffentlicht auf dem Publikationsserver der RWTH Aachen University
N1  - Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2025
AB  - As the demand for automation for mining industries increases, the need for autonomous transport vehicles becomes increasingly important. Articulated dump trucks (ADT), one of the main transport vehicles in mining operation, pose challenges due to their articulated steering mechanism. While considerable progress has been made in autonomous vehicle research, the dynamic modeling and specific control methods for ADTs remain underexplored. This thesis investigates two main aspects in the deployment of ADT in mining environments: simulation and control. The simulation aspect focuses on the challenges of dynamic modeling, especially due to the unique steering mechanism in ADTs. Therefore, a dynamic model is initially developed that emphasizes the indirect incorporation of steering dynamics to preserve model differentiability. This indirect approach allows the model to be used in model-based algorithms, extending its application beyond simulation alone. Following the successful implementation of the simulation environment, this thesis investigates Model-Based Predictive Controller (MPC) to achieve trajectory following for ADTs. The goal is to develop control strategies that support the full operational routine of an ADT. Given that phase delays and system constraints are inherent in ADTs, the MPC approach offers a viable solution due to its ability to incorporate system models, including their delays. Furthermore, this thesis presents the models required for the MPC, covering various operational scenarios and accounting for system delays. These delays have been identified through extensive testing and system identification. However, the system identification process in this work relies on offline methods, meaning it does not adapt to changing conditions in real time. In addition, this thesis investigates the integration of sideslip angle estimation and compensation as an important factor influencing control performance, especially while cornering. The results show that sideslip compensation improves control performance in such maneuvers. This thesis also differentiates between full-sized and compact ADTs, emphasizing that the faster steering dynamics of compact ADTs require tailored system models. All proposed methods have been validated through simulations and experimental setups, confirming the effectiveness and adaptability of the developed control strategies for real-world mining applications.
LB  - PUB:(DE-HGF)11
DO  - DOI:10.18154/RWTH-2025-09517
UR  - https://publications.rwth-aachen.de/record/1021168
ER  -