The dissertation of MSc (Tech) Abdolreza Taheri addresses learning controllers that drive heavy-duty machinery through advanced data-driven techniques, specifically reinforcement learning (RL) and transfer learning (TL). As industries strive for greater efficiency and sustainability, this research offers practical solutions that can lead to significant cost savings and reduced energy consumption in industrial hydraulic machines like loader cranes and wheel loaders.
The study explores four key research questions, demonstrating how RL can enhance control systems for hydraulic machines by developing a fast and scalable framework that “learns by interacting with a model of the machine”. It also tackles the challenges of modeling hydraulic actuators from real sensor measurements, ensuring applicability and reliable operation in real-world conditions. Additionally, the research highlights the potential for energy-efficient control by integrating pressure models into the RL training framework, which is crucial for minimizing environmental impact.
High-level automation solutions for industrial machines
The dissertation demonstrates high-level automation solutions for industrial machines, showcasing improved performance over classical control methods that are considered the industry standard.
“With four publications accompanying the dissertation, this work contributes to the field of industrial automation, providing insights that will shape the future of machine control systems and promote more sustainable industrial practices,” Abdolreza Taheri says.
Abdolreza received his MSc degree in Space Engineering from Sharif University of Technology, Iran in 2020. He joined Tampere University as a researcher in Marie Curie EU funded programme MORE-ITN, where he conducted the majority of his dissertation research in Sweden at companies HIAB and Volvo Construction Equipment (CE). After concluding his studies in 2023, he joined Volvo Group in Gothenburg, Sweden as a Data Scientist, working on scalable data-driven simulation platforms.
Public defence on Friday 6 June
The doctoral dissertation of MSc (Tech) Abdolreza Taheri, titled Transfer Learning for Real-world Control of Heavy-duty Hydraulic Machines, will be publicly examined at the Faculty of Engineering and Natural Sciences at Tampere University. The public defence takes place on Friday 6 June at 11 o’clock at Hervanta campus, Tietotalo building, in the auditorium TB104 (Korkeakoulunkatu 1, Tampere). The Opponent will be Associate Professor Johannes Stork from Örebro University, Sweden. The Custos will be Professor Reza Ghabcheloo from Tampere University.
The doctoral dissertation is available online.
The public defence can be followed via remote connection.