ARC@ORU: Programmatic Intelligence of Robotic Development and Learning

ARC@ORU Research Seminar Series

Programmatic Intelligence of Robotic Development and Learning

About the talk

Developing and training robotic platforms from the ground up is a time-consuming and difficult undertaking. Several techniques from different fields need to interact: (i) CAD modeling, 3D printing, and assembling of physical parts, (ii) physical modeling, system identification, and simulation of interacting components, and (iii) reinforcement learning and control with adequate software interaction. A fundamental question explored in our research group is how to co-design these parts such that composed systems can be designed and developed rapidly and correctly. In particular, we are interested in the combined problem of sim2real and explainable physical actuation. That is, how to construct reliable systems that can (i) transfer from simulated physical dynamics to real-world interaction, (ii) take action from ambiguous natural language instructions using large language models, and (iii) execute or reject tasks using a human-explainable approach. In this talk, I will discuss ongoing work addressing these problems. Especially, showcasing the development of a series of co-designed robotic platforms developed from scratch, including a new humanoid robot with 24 degrees of freedom. Besides the practical aspects of the design, we will discuss an approach called artificial mental models, the use of program language theory to control and explainability, as well as a new reinforcement learning technique for remotely controlling embedded systems with random communication latencies.

Speaker

David Broman is a Professor at the School of Electrical Engineering and Computer Science (EECS), KTH Royal Institute of Technology in Sweden. He is the Head of Department for the Department of Computing and Learning Systems (CLS) at KTH, and an Associate Director Faculty for the research center Digital Futures. He received his Ph.D. in Computer Science in 2010 from Linköping University, Sweden. Between 2012 and 2014, he was a Visiting Scholar at the University of California, Berkeley, where he was employed as a part-time researcher until 2016. Between 2023 and 2024, he was a Visiting Professor for a year at the Computer Science Department, Stanford University. His research concerns software and computing in general, with a focus on the intersection of (i) programming languages and compilers, (ii) probabilistic machine learning, and (iii) real-time and cyber-physical systems. David has received the Best ETAPS paper award on on programming languages and systems (the EAPLS Award, co-authored 2023), a Distinguished Artifact Award at ESOP (co-authored 2022), an outstanding paper award at RTAS (co-authored 2018), a best paper award in the journal Software & Systems Modeling (SoSyM award 2018), the award as teacher of the year, selected by the student union at KTH (2017), the best paper award at IoTDI (co-authored 2017), and awarded the Swedish Foundation for Strategic Research's individual grant for future research leaders (2016). He has worked for several years within the software industry, co-founded companies, co-founded the EOOLT workshop series, is the vice chair of the EMSOFT Steering Committee, is a member of IFIP Working group 2.4, the Modelica Association, the SWEDSOFT Board, a senior member of IEEE, and a former board member of Forskning och Framsteg.