Digit@LTH breakfast seminar: Modelling Intelligent Robot behaviour with Behavioural Trees by Volker Krueger
Place: E-huset, Ole Römers väg 3. Start in EIT Lunch-room (E:2328)
Contact: Jonas [dot] Wisbrant [at] cs [dot] lth [dot] se
Save event to your calendar
Title: Modelling Intelligent Robot behaviour with Behavioural Trees
Speaker: Professor Volker Krueger, Department of Computer Science
Location: M-house, Ole Römers väg 1. M:E
Abstract: Behavioral Trees (BT) have been used for a long time in computer games to give artificial characters a certain level of intelligence: With BTs the programmer can easily describe very complex behaviour patterns of the artificial game character, i.e., what the character should do in a given situation or context. This correlates very much with the problems in robotics where modern robots are expected to handle tasks in very different conditions and contexts.
Traditionally, finite state machines were used in robotics to describe complex behavioural patterns, the use of BTs is relatively new in the robotics community. In my talk, I want to discuss what BTs are, their definition, how they are constructed and how they are used. Then, in order to get a deeper understanding, I want to put the BTs into the context of FSM: Finite state machines (FSMs) are constructs from Theoretical computer science. They are well known from the area of formal language. We know they model the class of regular languages and we can related each FSM directly to a regular expression and vice versa. The question is: can BTs be analysed in the same way as FSMs? Can we also express a BT in form of a formal language? Are BTs even equivalent to FSMs?
Bio: Volker Krueger has studied computer science at the University of Kiel, Germany where he graduated with a Diploma degree (M.Sc.) in 1995. Dr. Krueger has completed his Dr. Ing. (PhD) in 2000 at Kiel University in the area of computer vision with a thesis of Gabor Wavelet Networks. He spent is PostDoc time in the Lab of Azriel Rosenfeld and Rama Chellappa at CFAR, University of Maryland before joining Aalborg University (DK) in 2002 as an Associate Professor. Until 2005 Dr. Krueger was researching in the area of Biometrics, Face recognition and Gait recognition. Since 2005, he has focused on cognitive robot in general and since 2012 on cognitive robotics particularly for manufacturing. He became Full Professor at Aalborg University in 2014. He was PI in the EU projects Paco-Plus(FP6), Tapas(FP7) and Scalable (H2020), and he was coordinator of the projects GISA (ECHORD/FP7) and STAMINA (FP7). Dr. Krueger has joined LTH in August 2018 as WASP professor for Autonomous Systems.
Registration: Please register no later than 27 March at 12.00 at https://www.lth.se/digitalth/events/register0328/