@misc{BestleHajipour, author = {Bestle, Dieter and Hajipour, Sanam}, title = {Design of a tracking controller based on machine learning}, series = {International journal of mechanical system dynamics}, volume = {5 / 2025}, journal = {International journal of mechanical system dynamics}, number = {2}, publisher = {Wiley}, address = {Hoboken, NJ}, issn = {2767-1402}, doi = {10.1002/msd2.70006}, pages = {201 -- 211}, abstract = {Tracking control of multibody systems is a challenging task requiring detailed modeling and control expertise. Especially in the case of closed-loop mechanisms, inverse kinematics as part of the controller may become a game stopper due to the extensive calculations required for solving nonlinear equations and inverting complicated functions. The procedure introduced in this paper substitutes such advanced human expertise by artificial intelligence through the utilization of surrogates, which may be trained from data obtained by classical simulation. The necessary steps are demonstrated along a parallel mechanism called λ-robot. Based on its mechanical model, the workspace is investigated, which is required to set proper initial conditions for generating data covering the used operation space of the robot. Based on these data, artificial neural networks are trained as surrogates for inverse kinematics and inverse dynamics. They provide forward control information such that the remaining error behavior is governed by a linear ordinary differential equation, which allows applying a linear quadratic regulator (LQR) from linear control theory. An additional feedback loop of the tracking error accounts for model uncertainties. Simulation results validate the applicability of the proposed concept.}, language = {en} } @misc{MammenKayatasBestle, author = {Mammen, Manasa Mariam and Kayatas, Zafer and Bestle, Dieter}, title = {Generation of multiple types of driving scenarios with variational autoencoders for autonomous driving}, series = {Future transportation}, volume = {5}, journal = {Future transportation}, number = {4}, publisher = {MDPI}, address = {Basel}, issn = {673-7590}, doi = {10.3390/futuretransp5040159}, pages = {1 -- 20}, abstract = {Generating realistic and diverse driving scenarios is essential for effective scenario-based testing and validation in autonomous driving and the development of driver assistance systems. Traditionally, parametric models are used as standard approaches for scenario generation, but they require detailed domain expertise, suffer from scalability issues, and often introduce biases due to idealizations. Recent research has demonstrated that AI models can generate more realistic driving scenarios with reduced manual effort. However, these models typically focused on single scenario types, such as cut-in maneuvers, which limits their applicability to diverse real-world driving situations. This paper, therefore, proposes a unified generative framework that can simultaneously generate multiple types of driving scenarios, including cut-in, cut-out, and cut-through maneuvers from both directions, thus covering six distinct driving behaviors. The model not only learns to generate realistic trajectories but also reflects the same statistical properties as observed in real-world data, which is essential for risk assessment. Comprehensive evaluations, including quantitative metrics and visualizations from detailed latent and physical space analyses, demonstrate that the unified model achieves comparable performance to individually trained models. The shown approach reduces modeling complexity and offers a scalable solution for generating diverse, safety-relevant driving scenarios, supporting robust testing and validation.}, language = {en} } @misc{ZhaiBestle, author = {Zhai, Tianyu and Bestle, Dieter}, title = {Design of nonlinear model predictive control for planar direct internal reforming solid oxide fuel cell stacks}, series = {2025 44th Chinese Control Conference (CCC)}, journal = {2025 44th Chinese Control Conference (CCC)}, publisher = {IEEE}, address = {Piscataway, NJ}, isbn = {978-988-75816-1-1}, doi = {10.23919/CCC64809.2025.11179672}, pages = {3093 -- 3098}, abstract = {The paper proposes an optimization-based nonlinear model predictive control (NMPC) strategy for the operation of a planar direct internal reforming solid oxide fuel cell (DIR-SOFC). The strategy focuses on regulating the inlet fuel and air flowrates to meet power demands, maintain the design temperature, and adhere to other operational constraints. A state estimator and optimizer are employed. Control and prediction horizons, as well as the time step lengths are carefully selected to balance estimation accuracy, stability, and computational cost. The obtained results demonstrate the effectiveness of NMPC in managing large-scale energy supply systems. Comparative analyses with varying tuning weights demonstrate the importance of appropriate weight selection regarding control performance and the robustness of NMPC. Both the dynamic modeling of the planar DIR-SOFC and the NMPC framework are implemented in the MATLAB/Simulink environment.}, language = {en} } @misc{HajipourZhaiBestle, author = {Hajipour, Sanam and Zhai, Tianyu and Bestle, Dieter}, title = {Data-driven sliding mode control of a rarallel 2D-robot}, series = {2025 8th International Conference on Intelligent Robotics and Control Engineering (IRCE)}, journal = {2025 8th International Conference on Intelligent Robotics and Control Engineering (IRCE)}, publisher = {IEEE}, address = {Piscataway, NJ}, isbn = {979-8-3503-5710-3}, doi = {10.1109/irce66030.2025.11203089}, pages = {75 -- 80}, abstract = {The paper introduces an easy-to-implement control strategy for a parallel 2 D robot. To avoid any analytic computations of multibody system dynamics and kinematics of the closed-loop mechanism, two feedforward networks are utilized as surrogates for inverse models. The necessary data are generated using classical simulation starting from random initial conditions and applying random force to cover the full state space used during control. Closed-loop stability of the lambda robot is guaranteed by extending the forward control based on inverse dynamics by a sliding mode controller with super-twisting structure. This approach results in an error decay to very low magnitude orders as an investigation of the tracking performance demonstrating high efficiency and control accuracy even when the plant parameters are disturbed.}, language = {en} } @misc{HajipourBestle, author = {Hajipour, Sanam and Bestle, Dieter}, title = {Surrogate-based robust tracking controller for a Lambda robot}, series = {Applied mathematics and mechanics : special issue : 95th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM)}, volume = {25}, journal = {Applied mathematics and mechanics : special issue : 95th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM)}, number = {3}, publisher = {Wiley-VCH}, address = {Weinheim}, doi = {10.1002/pamm.70017}, pages = {1 -- 8}, abstract = {Tracking control of multibody systems typically requires precise inversion of complex kinematic and dynamic equations, which can be particularly challenging for closed-loop mechanisms. This study presents a surrogate-based control approach that leverages artificial neural networks to approximate inverse kinematics and dynamics using data generated from classical simulations of a nominal model. These data-driven surrogates enable feedforward prediction of joint trajectories and control inputs without requiring analytical model inversion. To ensure robustness under modeling errors and parametric uncertainties, additional inner- and outer-loop sliding mode control (SMC) loops are integrated. Simulation results from a lambda robot demonstrate high tracking accuracy with asymptotic steady-state errors even under perturbations of system parameters. The proposed framework establishes a structured approach for trajectory tracking by combining data-driven feedforward and robust feedback control suitable for complex nonlinear robotic systems.}, language = {en} }