|December 13th: The Return of the Light|
|The First Örebro Winter School on "Artificial Intelligence and Robotics"|
The aim of this full-day course is to sketch an introduction to the use of knowledge representation and reasoning (KRR) techniques in robotics. The lectures will first provide a brief overview of the KRR techniques that may be relevant to the design and implementation of robotic systems. Then, the use of an explicit representation of knowledge will be exemplified through a case study involving semantic mapping (i.e. the construction of a map that embodies symbolic knowledge about the robot operational environment). The process for acquiring the knowledge about the environment and for building the symbolic representation of the semantic map will be illustrated in the lecture. The resulting representation will then be used by the students in the practicals to implement simple forms of reasoning in logic programming.
Prerequisites: basic programming capabilities in prolog and ROS programming (C++/Python) (in particular using ROSbuild and Groovy).
Daniele Nardi is full Professor at Sapienza Univ. Roma, Facoltà Ingegneria dell'Informazione, Informatica, Statistica, Dept. Computer, Control and Management Engineering "A. Ruberti". Laurea in Electrical Engineering, Politecnico Torino, 1981, Master Computer and System Engineering, Sapienza Univ. Roma, 1984. Researcher at Sapienza since 1988 (full professor since 2000). Recipient of "IJCAI-91 Publisher's Prize", prize "Intelligenza Artificiale 1993", ECCAI Fellow, President of RoboCup Federation (2011). Head of the research laboratory "Cognitive Robot Teams" (http:// http://www.dis.uniroma1.it/~nardi/), addressing different research topics: Cognitive Robotics, Localization, Navigation, Perception, Cooperation in multi-robot systems, Human Robot Interaction, Multimodal Interfaces and several application domains: Ambient Intelligence and robots to support elderly people, Disaster Response Robots to explore and gather information from the environment, Soccer Player robots for RoboCup competitions.
Guglielmo Gemignani earned his master's degree in Electronics Physics from Sapienza - Università di Roma cum laude and under the excellence program, after completing his bachelor's degree in Physics at Università degli Studi di Pisa. Currently Guglielmo is a PhD student at the Department of Computer, Control, and Management Engineering "Antonio Ruberti", Sapienza - Università di Roma and a member of the Ro.Co.Co. Laboratory (Cognitive and Cooperative Robots), held by Prof. Daniele Nardi. His PhD research program is mainly focused on Cognitive Robotics, Machine Learning, and Knowledge Representation and Reasoning. Under the supervision of Prof. Daniele Nardi, Guglielmo is currently working on the development of a cognitive agent able to learn and intelligently interact with multiple users in a domestic environment. Before starting to work at this project, Guglielmo has been involved in the SPQR RoboCup team, being in charge of the behaviours and coordination modules of the robotic team, as well as being one of the two team leaders during the whole 2012/13 academic year.Contact: http://www.dis.uniroma1.it/~nardi/
The technical contents of this full-day course will explore a specific automated planning paradigm which can reason about metric time and resources, and comprehensively deal with planning and scheduling of discrete tasks. The approach has demonstrated its applicability in synthesizing complex plans and executing them in real-world robot applications, including marine robots, smart homes and space applications. Constraint-based Planning (CBP) leverages its unique representational paradigm from earlier work in Control Theory and uses the notion of state variables and tokens within a flexible Simple Temporal Network representation which is ideal for embedded robotic platforms. The CBP approach we will discuss has flown in two NASA missions, the first in space (New Millennium Deep Space 1: Remote Agent [RAX]) and the first to control a vehicle on another planet (2004 Mars Exploration Rovers [MAPGEN]). More recently, the only operational closed loop AI-based system on an autonomous underwater vehicle was developed at MBARI with CBP techniques, and in Örebro a CBP approach was used to realize closed-loop, context-aware planning in a smart home with ubiquitous sensors and actuators.
Through the lectures and two separate sets of hands-on programming exercises, the course will provide a broad motivation for the representation formalism and its impact on the science of robotic autonomy.
Prerequisites: Students are expected to have gone thru their reading assignments prior to the event at Örebro.
Kanna Rajan is the Principal Researcher for Autonomy at the
Monterey Bay Aquarium Research Institute where he is engaged in autonomy
for marine robotics in the only AI group in operational
oceanography. Prior to coming to MBARI in 2005 he spent a decade at the
NASA Ames Research Center where he was on two NASA missions, the New
Millennium Deep Space 1 Remote Agent Experiment in 1999 and the 2003
Mars Exploration Rovers mission, the former as a part of the AI Planner
team, the latter as the Principal Investigator of the MAPGEN system, the
longest running AI system which continues to command the Opportunity
rover from earth. He was in the doctoral program at NYU/Courant
Institute prior to which he was in the Knowledge Systems group at
American Airlines. He was the co-Chair of the 2005 Intnl. Conference on
Automated Planning and Scheduling (ICAPS) and his academic interests lie
in autonomy architectures and representation for real-world
environments. His recent publications can be found at:
Frédéric Py received the Ph.D. in robotic architectures from the University Paul Sabatier at the Laboratory for Analysis and Architecture of Systems (LAAS-CNRS), Toulouse, France, in 2005. He worked as a Software Engineer in the Autonomous Systems Group, Monterey Bay Aquarium Research Institute (MBARI), Moss Landing, CA. Were he took over the design of the T-REX architecture supporting constraint-based planning and plan execution and maintained it on MBARI's Dorado AUV. His focus is still on T-REX architecture and more precisely how to improve the interaction between planning and execution in a situated agent.
Federico Pecora (PhD in Computer Science from University of Rome "La Sapienza", 2007) is a Senior Lecturer at the Center for Applied Autonomous Sensor Systems (AASS) at Örebro University, Sweden. His interests lie at the intersection of Artificial Intelligence and Robotics, with a focus temporal and spatial constraint reasoning algorithms, planning, scheduling, meta-CSP techniques for integrated reasoning, constraint-based planning and scheduling for robotic systems, and context recognition. He has primarily applied these techniques in two broad application areas: service robots/sensor systems for use in domestic environments; and decision support tools for industrial scenarios with large autonomous vehicles.Contact: firstname.lastname@example.org
The rise of the World Wide Web, the mobile Web and the Internet of Things will likely have massive impact on robotics: Web sites like ehow.com or epicurious.com provide instructions for millions of everyday tasks and cooking recipes, image search engines or repositories like the Trimble/Google 3D warehouse contain object pictures and 3D models, shopping websites and product catalogues contribute semantic information about objects. Cloud Computing services like Google Goggles for object recognition can help robots to identify millions of different objects, as needed in the open world. Similar techniques could also be used to offload computation- or storage-intensive robot capabilities such as mapping to the Cloud.
This tutorial will cover different ways in which robots can use the Internet: Using textual information from the Web that was created by and for humans, mining object models from Web resources, and using cloud services specifically developed for robots.
Prerequisites: Students are expected to have gone thru their reading assignments prior to the event at Örebro.
Moritz Tenorth is currently a post-doc in the Institute for Artificial Intelligence at the University of Bremen, Germany. His research interests include grounded knowledge representations which integrate information from web sources, observed sensor data and data mining techniques, and their applications to knowledge-based action interpretation and robot control. He received his doctoral degree in 2011 from Technische Universität München, Germany, after studying electical engineering in Aachen and Paris, obtaining his Diploma degree (equivalent to M.Eng.) in 2007 from RWTH Aachen, Germany.Contact: http://ai.uni-bremen.de/team/moritz_tenorth/
Do you know - now, as you started reading this text - where exactly the remote control of your TV set is? And your high-school and university diplomas? Salt jar and sugar bowl? For some of these, you may have firm beliefs: The diplomas, for example, may be safely stowed away in a constant place. For others, matters are different. There may be a nominal place for the remote control, but most of us know that we don¢t know where it is unless we see it. Alas, your huge personal knowledge base has a gap, probably even many, and you are aware of it. Does that prevent you from reading this abstract now? Obviously not - at least if the gap concerns only the remote control at home. It certainly would if you were to realize that your laptop or smartphone to read this were suddenly gone. The difference is: You need the laptop/smartphone now; but the remote control is currently irrelevant, judging from what you are doing right now.
Informally, one might say the laptop is currently in your knowledge focus, the remote control is not. You need not care about gaps that are out of focus, i.e., of no concern for what you are doing or about to do. The same goes for contradictions: You may have conflicting information about your personal program for tonight, but you need not care about it right now. Very likely, the ability of not caring about knowledge out of focus is a pre-requisite for being ... well, focused on some particular task.
Without pushing the analogy to biological cognition any further, this course deals on a conceptual and technical level with focus, in the above sense, in robot control. We will argue in this course that the effort to develop methods for setting up and maintaining a focus on a subset of a robot¢s knowledge base for its reasoning is worth it, if not necessary; and we will sketch how to attempt it.
Joachim Hertzberg is a full professor for computer science at Osnabrück University since 2004, heading the Knowledge-Based Systems group. With a secondary affiliation, he is heading the Osnabrück Branch of DFKI's (German Research Center for Artificial Intelligence) Robotics Innovation Center. He has graduated in Computer Science (U. Bonn, 1982; Dr.rer.nat. 1986, U. Bonn; habilitation 1995, U. Hamburg). Former affiliations include Bonn University and the institutes GMD and Fraunhofer AIS in Sankt Augustin, Germany. His areas of research are AI and Mobile Robotics, with contributions to action planning, plan-based robot control, sensor data interpretation, semantic mapping, embedded knowledge-based systems, reasoning about action, constraint-based reasoning, and applications of these.Contact: http://www.inf.uos.de/hertzberg/
Human-robot interaction requires to equip the robot with explicit reasoning on the human and on its own capacities to achieve its tasks in a collaborative way with a human partner. This talk presents a robot control system which has been especially designed for a cognitive robot which shares space and task with a human. We have adopted a constructive approach based on effective individual and collaborative skills. The system is comprehensive since it aims at dealing with a complete set of abilities articulated so that the robot controller is effectively able to conduct a collaborative task with a human partner in a flexible manner
These abilities include geometric reasoning and situation assessment based essentially on perspective-taking and affordances, management and exploitation of each agent (human and robot) knowledge in a separate cognitive model, human-aware task planning and human and robot interleaved plan achievement.
I will present and discuss these different topics and give concrete instances of their use.
Rachid Alami is Senior Scientist at CNRS. He received an engineer diploma in computer science in 1978 from ENSEEIHT, and a Ph.D in Robotics in 1983 from the University Paul Sabatier. He contributed and took important responsibilities in several national, European and international research and/or collaborative projects (EUREKA: FAMOS, AMR and I-ARES projects, ESPRIT: MARTHA, PROMotion, ECLA, IST: COMETS, IST FP6 and FP7 projects COGNIRON, URUS, PHRIENDS, CHRIS, SAPHARI, ARCAS, SPENCER; France: ARA, VAP-RISP for planetary rovers, PROMIP, ANR projects). His main research contributions fall in the fields of Robot Architectures, Task and motion planning, multi-robot cooperation, and human-robot interaction. Rachid Alami is currently the head of the Robotics and AI Department at LAAS.Contact: http://homepages.laas.fr/rachid/
This lecture will provide an overview of artificial cognition, focussing on the link with robotics. After an initial discussion of the capabilities of a cognitive system - goal directed action, perception, learning, anticipation, and adaptation - we will explain different cognitive architectures before proceeding to discuss in more detail the complex issues of autonomy and embodiment. We will wrap up the lecture by reflecting on the importance of development and the role of social cognition in human robot interaction.
David Vernon is a Professor of Informatics at the University of Skövde, Sweden. He works on cognitive systems and is particularly interested in modelling autonomy. Over the past 34 years, he has held positions at Westinghouse Electric, Trinity College Dublin, the European Commission, the National University of Ireland, Maynooth, Science Foundation Ireland, Khalifa University, UAE, and the Technical University of Munich.Contact: http://www.vernon.eu
[ School Home ]