Download a PDF with the full list of Robot-Intelligence-Lab-Publications.pdf

A comprehensive list can also be found at Google Scholar, or by searching for the publications of author Kormushev, Petar.

Search or filter publications

Filter by type:

Filter by publication type

Filter by year:

to

Results

  • Showing results for:
  • Reset all filters

Search results

  • CONFERENCE PAPER
    Kormushev P, Calinon S, Caldwell DG, 2010,

    Approaches for Learning Human-like Motor Skills which Require Variable Stiffness During Execution

    , IEEE Intl Conf. on Humanoid Robots (Humanoids), Workshop on Humanoid Robots Learning from Human Interaction

    Humans employ varying stiffness in everydaylife for almost all human motor skills, using both passiveand active compliance. Robots have only recently acquiredvariable passive stiffness actuators and they are not yet mature.Active compliance controllers have existed for a longer time,but the problem of automatic determination of the necessarycompliance to achieve a task has not been thoroughly studied.Teaching humanoid robots to apply variable stiffness to theskills they acquire is vital in order to achieve human-likenaturalness of the execution. Also, using adaptive compliancecan help to increase the energy efficiency. This paper comparestwo different approaches that allow robots to learn human-likeskills which require varying stiffness during execution. Theadvantages and disadvantages of each approach is discussedand demonstrated with various experiments on an activelycompliantBarrett WAM robot.

  • CONFERENCE PAPER
    Kormushev P, Calinon S, Saegusa R, Metta Get al., 2010,

    Learning the skill of archery by a humanoid robot iCub

    , Pages: 417-423

    We present an integrated approach allowing the humanoid robot iCub to learn the skill of archery. After being instructed how to hold the bow and release the arrow, the robot learns by itself to shoot the arrow in such a way that it hits the center of the target. Two learning algorithms are proposed and compared to learn the bi-manual skill: one with Expectation-Maximization based Reinforcement Learning, and one with chained vector regression called the ARCHER algorithm. Both algorithms are used to modulate and coordinate the motion of the two hands, while an inverse kinematics controller is used for the motion of the arms. The image processing part recognizes where the arrow hits the target and is based on Gaussian Mixture Models for color-based detection of the target and the arrow's tip. The approach is evaluated on a 53-DOF humanoid robot iCub. ©2010 IEEE.

  • CONFERENCE PAPER
    Sato F, Nishii T, Takahashi J, Yoshida Y, Mitsuhashi M, Kormushev P, Kanamiya Yet al., 2010,

    Whiteboard Cleaning Task Realization with HOAP-2

    , 2010 IEEE/SICE International Symposium on System Integration SII 2010, Pages: 426-429

    The controller has a motion feedback control loop for the four arm joints and a force feedback controlloop for the ankle joint. The desired motion/force trajectories are obtained via kinesthetic teaching by a human,followed by a learning phase. During skill demonstration, the arm joints are made compliant while the robotbalances using ankle/hip strategies. Experimental data demonstrate the feasibility of the method.

  • CONFERENCE PAPER
    Kormushev P, Dong F, Hirota K, 2009,

    Probability redistribution using time hopping for reinforcement learning

    , 10th International Symposium on Advanced Intelligent Systems, ISIS 2009

    —A method for using the Time Hopping techniqueas a tool for probability redistribution is proposed. Applied toreinforcement learning in a simulation, it is able to re-shape thestate probability distribution of the underlying Markov decisionprocess as desired. This is achieved by modifying the targetselection strategy of Time Hopping appropriately. Experimentswith a robot maze reinforcement learning problem show that themethod improves the exploration efficiency by re-shaping thestate probability distribution to an almost uniform distribution.

  • JOURNAL ARTICLE
    Kormushev P, Nomoto K, Dong F, Hirota Ket al., 2008,

    Time manipulation technique for speeding up reinforcement learning in simulations

    , Cybernetics and Information Technologies, Vol: 8, Pages: 12-24, ISSN: 1311-9702

    A technique for speeding up reinforcement learning algorithms by usingtime manipulation is proposed. It is applicable to failure-avoidance controlproblems running in a computer simulation. Turning the time of the simulationbackwards on failure events is shown to speed up the learning by 260% andimprove the state space exploration by 12% on the cart-pole balancing task,compared to the conventional Q-learning and Actor-Critic algorithms.

  • CONFERENCE PAPER
    Yamazaki Y, Dong F, Masuda Y, Uehara Y, Kormushev P, Vu HA, Le PQ, Hirota Ket al., 2007,

    Fuzzy inference based mentality estimation for eye robot agent

    , 23rd Fuzzy System Symposium, Pages: 387-388

    Household robots need to communicate with human beings in a friendly fashion. To achievebetter understanding of displayed information, an importance and a certainty of the information shouldbe communicated together with the main information. The proposed intent expression system aims toconvey this additional information using an eye robot. The eye motions are represented as states in apleasure-arousal space model. Change of the model state is calculated by fuzzy inference according to theimportance and certainty of the displayed information. This change influences the arousal-sleepcoordinate in the space which corresponds to activeness in communication. The eye robot provides abasic interface for the mascot robot system which is an easy to understand information terminal forhome environments in a humatronics society.

  • CONFERENCE PAPER
    Yamazaki Y, Dong F, Masuda Y, Uehara Y, Kormushev P, Vu HA, Le PQ, Hirota Ket al., 2007,

    Intent expression using eye robot for mascot robot system

    An intent expression system using eye robots is proposed for a mascot robotsystem from a viewpoint of humatronics. The eye robot aims at providing a basicinterface method for an information terminal robot system. To achieve betterunderstanding of the displayed information, the importance and the degree ofcertainty of the information should be communicated along with the maincontent. The proposed intent expression system aims at conveying thisadditional information using the eye robot system. Eye motions are representedas the states in a pleasure-arousal space model. Changes in the model state arecalculated by fuzzy inference according to the importance and degree ofcertainty of the displayed information. These changes influence thearousal-sleep coordinates in the space that corresponds to levels of livelinessduring communication. The eye robot provides a basic interface for the mascotrobot system that is easy to be understood as an information terminal for homeenvironments in a humatronics society.

  • JOURNAL ARTICLE
    Agre G, Kormushev P, Dilov I, 2006,

    INFRAWEBS axiom editor - a graphical ontology-driven tool for creating complex logical expressions

    , International Journal of Information Theories and Applications, Vol: 13, Pages: 169-178

    The current INFRAWEBS European research project aims at developing ICTframework enabling software and service providers to generate and establishopen and extensible development platforms for Web Service applications. One ofthe concrete project objectives is developing a full-life-cycle softwaretoolset for creating and maintaining Semantic Web Services (SWSs) supportingspecific applications based on Web Service Modelling Ontology (WSMO) framework.According to WSMO, functional and behavioural descriptions of a SWS may berepresented by means of complex logical expressions (axioms). The paperdescribes a specialized user-friendly tool for constructing and editing suchaxioms - INFRAWEBS Axiom Editor. After discussing the main design principles ofthe Editor, its functional architecture is briefly presented. The tool isimplemented in Eclipse Graphical Environment Framework and Eclipse Rich ClientPlatform.

  • CONFERENCE PAPER
    Agre G, Kormushev P, Dilov I, 2005,

    INFRAWEBS Capability Editor - A graphical ontology-driven tool for creating capabilities of Semantic Web Services

    , Third International Conference Information Research, Applications and Education, Pages: 228-228

    The current INFRAWEBS European research project aims at developing ICT frameworkenabling software and service providers to generate and establish open and extensible development platforms forWeb Service applications. One of the concrete project objectives is developing a full-life-cycle software toolset forcreating and maintaining Semantic Web Services (SWSs) supporting specific applications based on Web ServiceModelling Ontology (WSMO) framework. SWSs are self-contained, self-describing, semantically marked-upsoftware resources that can be published, discovered, composed and executed across the Web in a task drivensemi-automatic way. A main part of WSMO-based SWS is service capability – a declarative description of Webservice functionality. A formal syntax and semantics for such a description is provided by Web Service ModelingLanguage (WSML), which is based on different logical formalisms, namely, Description Logics, First-Order Logicand Logic Programming. A WSML description of a Web service capability is represented as a set of complexlogical expressions (axioms). The paper describes a specialized user-friendly tool for constructing and editingWSMO-based SWS capabilities. Since the users of that tool are assumed to be SWS providers, which are not thespecialists in first-order logic, it is proposed a graphical way for constricting and editing the axioms abstractingaway as much as possible from a concrete syntax of logical language used for implementing them.Our analysis has shown that the main problems arising during axiom creating are associated with using of correctnames of concepts, attributes, relations and parameters as well as their types rather than with expressing logicaldependences between axiom parts. So the process of constructing logical expressions in the tool is ontologydriven,which means that in each step of this process the user may select only such elements of existingontologies that are consistent with already constructed part of the axiom.

This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.

Request URL: http://wlsprd.imperial.ac.uk:80/respub/WEB-INF/jsp/search-t4-html.jsp Request URI: /respub/WEB-INF/jsp/search-t4-html.jsp Query String: id=815&limit=20&page=3&respub-action=search.html Current Millis: 1513409570857 Current Time: Sat Dec 16 07:32:50 GMT 2017