Simon Hirlaender

Julian Langschwert, Georg Schaefer, Jakob Rehrl et al.

Informative excitation signals are critical for accurate system identification of mechatronic systems, yet classical system identification (SI) approaches require expert knowledge and hand-crafted signal design to respect hardware safety constraints, limiting their generalizability. We propose a reinforcement learning (RL) agent that learns optimal excitation signals for a Quanser Aero 2 testbed while autonomously enforcing safety constraints through reward shaping. Evaluated across 10 independent training seeds, our comprehensive agent achieves competitive estimation accuracy across all three identified parameters, outperforming classical baselines while incurring only 0.75% safety violations.

Georg Schäfer, Jakob Rehrl, Stefan Huber et al.

This study conducts a comparative analysis of Model Predictive Control (MPC) and Proximal Policy Optimization (PPO), a Deep Reinforcement Learning (DRL) algorithm, applied to a 1-Degree of Freedom (DOF) Quanser Aero 2 system. Classical control techniques such as MPC and Linear Quadratic Regulator (LQR) are widely used due to their theoretical foundation and practical effectiveness. However, with advancements in computational techniques and machine learning, DRL approaches like PPO have gained traction in solving optimal control problems through environment interaction. This paper systematically evaluates the dynamic response characteristics of PPO and MPC, comparing their performance, computational resource consumption, and implementation complexity. Experimental results show that while LQR achieves the best steady-state accuracy, PPO excels in rise-time and adaptability, making it a promising approach for applications requiring rapid response and adaptability. Additionally, we have established a baseline for future RL-related research on this specific testbed. We also discuss the strengths and limitations of each control strategy, providing recommendations for selecting appropriate controllers for real-world scenarios.

Simon Hirlaender, Niky Bruchon

Reinforcement learning holds tremendous promise in accelerator controls. The primary goal of this paper is to show how this approach can be utilised on an operational level on accelerator physics problems. Despite the success of model-free reinforcement learning in several domains, sample-efficiency still is a bottle-neck, which might be encompassed by model-based methods. We compare well-suited purely model-based to model-free reinforcement learning applied to the intensity optimisation on the FERMI FEL system. We find that the model-based approach demonstrates higher representational power and sample-efficiency, while the asymptotic performance of the model-free method is slightly superior. The model-based algorithm is implemented in a DYNA-style using an uncertainty aware model, and the model-free algorithm is based on tailored deep Q-learning. In both cases, the algorithms were implemented in a way, which presents increased noise robustness as omnipresent in accelerator control problems. Code is released in https://github.com/MathPhysSim/FERMI_RL_Paper.

Georg Schäfer, Max Schirl, Jakob Rehrl et al.

This paper proposes a framework for training Reinforcement Learning agents using Python in conjunction with Simulink models. Leveraging Python's superior customization options and popular libraries like Stable Baselines3, we aim to bridge the gap between the established Simulink environment and the flexibility of Python for training bleeding edge agents. Our approach is demonstrated on the Quanser Aero 2, a versatile dual-rotor helicopter. We show that policies trained on Simulink models can be seamlessly transferred to the real system, enabling efficient development and deployment of Reinforcement Learning agents for control tasks. Through systematic integration steps, including C-code generation from Simulink, DLL compilation, and Python interface development, we establish a robust framework for training agents on Simulink models. Experimental results demonstrate the effectiveness of our approach, surpassing previous efforts and highlighting the potential of combining Simulink with Python for Reinforcement Learning research and applications.

Georg Schäfer, Tatjana Krau, Jakob Rehrl et al.

Reinforcement Learning (RL) offers promising solutions for control tasks in industrial cyber-physical systems (ICPSs), yet its real-world adoption remains limited. This paper demonstrates how seemingly small but well-designed modifications to the RL problem formulation can substantially improve performance, stability, and sample efficiency. We identify and investigate key elements of RL problem formulation and show that these enhance both learning speed and final policy quality. Our experiments use a one-degree-of-freedom (1-DoF) helicopter testbed, the Quanser Aero~2, which features non-linear dynamics representative of many industrial settings. In simulation, the proposed problem design principles yield more reliable and efficient training, and we further validate these results by training the agent directly on physical hardware. The encouraging real-world outcomes highlight the potential of RL for ICPS, especially when careful attention is paid to the design principles of problem formulation. Overall, our study underscores the crucial role of thoughtful problem formulation in bridging the gap between RL research and the demands of real-world industrial systems.

Georg Schäfer, Raphael Seliger, Jakob Rehrl et al.

Industrial automation increasingly demands energy-efficient control strategies to balance performance with environmental and cost constraints. In this work, we present a multi-objective reinforcement learning (MORL) framework for energy-efficient control of the Quanser Aero 2 testbed in its one-degree-of-freedom configuration. We design a composite reward function that simultaneously penalizes tracking error and electrical power consumption. Preliminary experiments explore the influence of varying the Energy penalty weight, alpha, on the trade-off between pitch tracking and energy savings. Our results reveal a marked performance shift for alpha values between 0.0 and 0.25, with non-Pareto optimal solutions emerging at lower alpha values, on both the simulation and the real system. We hypothesize that these effects may be attributed to artifacts introduced by the adaptive behavior of the Adam optimizer, which could bias the learning process and favor bang-bang control strategies. Future work will focus on automating alpha selection through Gaussian Process-based Pareto front modeling and transitioning the approach from simulation to real-world deployment.