Personal tools
You are here: Home Members Mario Prats Homepage

Mario Prats

Robotic Intelligence Lab, Interactive and Robotic Systems, Ph.D.
Contact    Academic background    Research     Publications    Press     Software

 

Contact

e-Mail: marioprats AT gmail.com
Office: TI-2034-DI,Robotic Intelligence Laboratory, Universitat Jaume I
Address: Avda. Vicente Sos Baynat s/n, E-12071, Castelló, Spain
Tel: +34 964 387 048
Fax: +34 964 728 486
Personal website:  http://www.robotic-creatures.com
Twitter:  marioprats
Google+:  Google+ profile
LinkedIn:  LinkedIn profile
YouTube:  My channel,  IRSLab channel

  

Academic background

  • Ph.D. by Jaume-I University (June 2009) entitled Robot Physical Interaction through the combination of Vision, Tactile and Force Feedback. Received EURON Georges Giralt Award 2011, Robotdalen International Scientific Award Honorable Mention 2009 and finalist of the CEA-GTRob award to the best PhD thesis on Robotics
  • Master Degree in Computer Science Engineering at Jaume-I University, Castellón (March 2004).
  • Computer Science Engineering at Jaume-I University, Castellón (2003)

Current research

Since October 2012 I'm working at Willow Garage, Menlo Park, CA.

Past research (TRIDENT FP7 Project)

In the period from 2010 to 2012 I worked on the development of autonomous manipulation systems for performing underwater tasks. The problem is very similar to the mobile manipulation problem, but normally with more degrees of freedom. In fact, an underwater vehicle typically has four DOF, but some of them can be controlled even on the six DOF. The degrees of freedom of the manipulator have to be added to these. In addition to the problem of controlling a high number of degrees of freedom, the vehicle cannot easily keep its pose while the manipulator is doing the task. On the contrary, the vehicle is a floating platform that tries to keep the pose as accurately as possible, but is highly affected by underwater currents, coupled dynamics, etc. This means that the manipulation task has to be normally performed under some motion of the base. Finally, vision also suffers from the underwater environment because of the bad visibility, floating particles, color attenuation, etc. All these facts make autonomous underwater manipulation a challenging problems that is worth researching. This research was carried out in the context of the RAUVI and TRIDENT projects. For more info take a look to my personal blog or my website at the Interactive and Robotic Systems group.

Past research

  • Master degree project developped at LASMEA/IFMA, in France, related to force/vision coupling (2002-2003)
  • Three-months stay at Intelligent Systems Research Center, Sungkyunkwan University, Suwon, South Korea, researching on force/vision control for robotic execution of daily tasks (2007).
  • Five-months stay in Karlsruhe, Germany, researching on the force control of a humanoid robot for performing everyday tasks, taking into account redundancy at joint and task level (2007).
  • Current research is being developped at Robotic Intelligence Lab, Universitat Jaume I, Castellón.

The purpose of our research is to advance towards autonomous, robust and versatile manipulation. More concretely, we support that a further integration of two aspects that have been normally considered independently, the grasp and the task, could imply a breakthrough in robotic manipulation.

We advocate for the introduction of task-related aspects into the classical knowledge-based grasp concept, leading to task-oriented grasps. In a similar manner, grasp-related issues are also considered during the execution of a task, leading to grasp-oriented tasks. Both task-oriented grasps and grasp-oriented tasks compose a unified representation which we call physical interaction. This new concept suppresses the classical boundaries between the grasp and the task, and introduces the new problems of physical interaction specification, planning and execution.

Regarding the specification of physical interaction tasks, we have defined a framework (published at MFI08) that allows to define the grasp and task-related aspects of a task in an integrated manner:

  

 

A task-oriented grasp planner has been developed in order to autonomously specify physical interaction tasks concerning manipulation of articulated objects and household devices. Objects are located and tracked in the environment using vision:

   
   
   
   

 

Regarding sensor-based execution of physical interaction tasks, different experiments have been carried out combining multiple sensors and using different robots. In the line of vision/force coupling, we developped a project during 6 months at LASMEA, in Clermont-Ferrand, France, under the supervision of professors Youcef Mezouar and Philippe Martinet. In that work, a manipulator robot was programmed in order to perform visual servoing tasks in a compliant manner, by means of a force sensor attached to the wrist. This work lead to a new approach for coupling vision and force feedback: the external hybrid vision-force control. Additional vision/force experiments were made at the Intelligent Systems Research Center (Sungkyunkwan University, South Korea), in collaboration with professors Philippe Martinet and Sukhan Lee. During this stay, a 7DOF robotic arm was programmed to perform door opening tasks by means of external vision/force control.

 

The robot was observing simultaneously his hand and the object to manipulate, by using an external camera (i.e. robot head). Task-oriented grasping algorithms were used in order to plan a suitable grasp on the object according to the task to perform. External vision/force control, was used in order to, first, guide the robot hand towards the grasp position and, second, perform the task taking into account external forces. The coupling between these two complementary sensor modalities provided the robot with robustness against uncertainties in models and positioning. A position-based visual servoing control law was designed in order to continously align the robot hand with respect to the object that was manipulated, independently of camera position. This allowed to freely move the camera while the task was executed, making this approach amenable to be integrated in current humanoid robots without the need of hand-eye calibration. More details can be found in the corresponding publication.

During a stay at Karlsruhe University (Germany), force control with joint redundancy management for service tasks, was implemented on the Armar-III Humanoid robot. A total of 8 degrees of freedom (7 in the arm, 1 in the waist) and a five-fingered hand were used for interacting with the different furniture found in a kitchen environment. The use of redundancy at the task level, as well as the joint-level redundancy of the robot, allowed us to have many redundant degrees of freedom that were controlled in order to avoid joint limits, following a secondary task approach. We are very grateful to Steven Wieland, Tamim Asfour and Prof. Dillmann for their support in this project. This work led to a publication at HUMANOIDS 2008


    

Armar-III SFB-588 Karlsruhe 

 

Some of the developments on vision/force control have also been applied to the UJI Librarian Robot, which is a robot able to autonomously search for a book in a library and take it to the user. The UJI Librarian Robot gave place to the UJI Service Robot, which incorporates a three-fingered hand that allows the robot to perform more natural movements for book grasping. Also, some results have been obtained in the field of autonomous manipulation for performing human daily tasks, as shown in the following images.

Further reliability in the book grasping capability was achieved by the installation of tactile sensors at the fingertips:

     

Current research is focused on the integration of force, vision and tactile sensors for reliable physical interaction. Force feedback is at the lowest level, ensuring a safe manipulation even if the rest of sensors provide innacurate information. Pose estimation methods based on the Virtual Visual Servoing approach have been developed in order to perform vision-based manipulation by tracking the natural object features instead of special markers. Tactile feedback allows to accurately control some degrees of freedom which can not be reliably controlled by force or vision. Some snapshots of different experiments are included here (more details in the ICRA09 paper):

    
     

Publications

 INTERNATIONAL JOURNALS

  • M. Prats, P.J. Sanz and A.P. del Pobil. Reliable non-prehensile door opening through the combination of vision, tactile and force feedback. Autonomous Robots, 29(2), pp. 201-218, 2010.
  • M. Prats, P.J. Sanz and A.P. del Pobil. A Framework for Compliant Physical Interaction: the grasp meets the task. Autonomous Robots, 28(1), pp. 89-111, 2010.
  • M. Prats, E. Martínez, P.J. Sanz, A.P. del Pobil. The UJI Librarian Robot. Journal of Intelligent Service Robotics, vol. 1, num. 4, pp 321-335, Oct. 2008.
  • M. Prats, P.J. Sanz, A.P del Pobil, E. Martínez, R. Marín. Towards multipurpose autonomous Manipulation with the UJI service robot. ROBOTICA, vol. 25, pp. 245-256, Cambridge University Press, 2007.
  • M. Prats, P. Martinet, A.P. del Pobil, S. Lee. Robotic Execution of Everyday Tasks by means of External Vision/Force Control. Journal of Intelligent Service Robotics, vol. 1, num. 3, pp. 253–266, 2008.

NATIONAL JOURNALS

  • M. Prats, P.J. Sanz, E. Martínez, R. Marín, A. P. del Pobil. Manipulación Autónoma Multipropósito en el Robot de Servicios Jaume-2. Revista Iberoamericana de Automática Informática Industrial, vol. 5, num. 2, Abril 2008, pp. 25-37. ISSN: 1697-7912 

BOOK CHAPTERS

  • M. Prats, P. Martinet, S. Lee, P.J. Sanz. Compliant Physical Interaction based on External Vision-Force Control and Tactile-Force Combination. Lecture Notes in Electrical Engineering , Vol. 35. Lee, S.; Ko, Hanseok; Hahn, Hernsoo (Eds.) 2009. ISBN: 978-3-540-89858-

INTERNATIONAL CONFERENCES

  • J.J Sorribes, M. Prats and A. Morales. Visual Robot Hand Tracking based on Articulated 3D Models for Grasping. Accepted for IEEE International Conference on Robotics and Automation, Anchorage, Alaska, 2010.
  • M. Prats, P.J. Sanz and A.P del Pobil. Vision-Tactile-Force Integration and Robot Physical Interaction. IEEE International Conference on Robotics and Automation, pp. 3975–3980, Kobe, Japan, 2009.
  • M. Prats, S. Wieland, T. Asfour, A.P. del Pobil and R. Dillmann. Compliant Interaction in Household Environments by the Armar-III Humanoid Robot. IEEE-RAS International Conference on Humanoid Robots, pp 475–480, Korea, 2008.
  • M. Prats, P. Martinet, S. Lee, P. J. Sanz. Compliant Physical Interaction based on External Vision-Force Control and Tactile-Force Combination. International Conference on Multisensor Fusion and Integration for Intelligent Systems, Seoul (Korea), 2008.
  • M. Prats, P. J. Sanz, A. P. del Pobil. A Framework for Compliant Physical Interaction based on Multisensor Information. International Conference on Multisensor Fusion and Integration for Intelligent Systems. Seoul (Korea), 2008.
  • M. Prats, P. Martinet, A. P. del Pobil, S. Lee. Vision/Force Control in Task-Oriented Grasping and Manipulation. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS’07), Oct 29 - Nov 2, Sheraton Hotel, San Diego, CA, USA, 2007.
  • M. Prats, P. J. Sanz, A. P. del Pobil. Task Planning for Intelligent Robot Manipulation. In Artificial Intelligence and Applications, Innsbruck, Austria February 12-14, 2007.
  • M. Prats, P. J. Sanz, A. P. del Pobil. Task-Oriented Grasping using Hand Preshapes and Task Frames. In IEEE International Conference on Robotics and Automation (ICRA’07), 10-14 April, Roma, Italy, 2007. 
  • A. Morales, M. Prats, G. Recatalá. A Framework for using Tactile Data in Autonomous Reactive Manipulation. In Robotics and Applications and Telematics (RA’07), 29-31 August, Würzburg, Germany, 2007
  • Youcef Mezouar, M. Prats, P. Martinet. External Hybrid Vision/Force Control. In International Conference on Advanced Robotics, August 21-24, Jeju Island, South Korea, 2007.
  • M. Prats, A. P. del Pobil, P. J. Sanz. A Control Architecture for Compliant Execution of Manipulation Tasks. In Proc. of International Conference on Intelligent Robots and Systems, Be¼ing, China, October 9-15. 2006.
  • M. Prats, P. J. Sanz, A. P. del Pobil. Model-based Tracking and Hybrid Force/Vision Control for the UJI Librarian Robot. In Proc. of International Conference on Intelligent Robots and Systems, Edmonton, Canada 2005.
  • E. Martínez, M. Prats, A. P. del Pobil, P. J. Sanz. Robots Behave in the Real World: Looking for Books in a Library. In the 9th IASTED International Conference on Artificial Intelligence and Soft Computing, 12-14 September, Benidorm, Spain 2005.
  • M. Prats, P. J. Sanz, A. P. del Pobil. Perception-Based Search and Manipulation in a Semi-Structured Environment. In Proc. of International Conference on Advanced Robotics (ICAR05), Seattle, USA. pp. 749 - 754, 2005.
  • M. Prats, R. Ramos-Gar¼o, P. J. Sanz, A. P. del Pobil. Autonomous Localization and Extraction of Books in a Library. In Intelligent Autonomous Systems, edited by F. Groen et al., IOS Press, Amsterdam 2004.
  • Prats M., Ramos-Gar¼o R., Sanz P.J., del Pobil A.P. A Robot for Intelligent Book Handling. Artificial Intelligence and Applications, Vol. 2, pp. 718–723, 2004.
  • R. Ramos-Garijo, M. Prats, P. J. Sanz, A. P. del Pobil. Recent Progress in the UJI Librarian Robot. In IEEE Int. Conference on Systems, Man & Cybernetics, The Hague, The Netherlands 2004.
  • R. Ramos-Garijo, M. Prats, P. J. Sanz, A. P. del Pobil. An Autonomous Assistant Robot for Book Manipulation in a Library. In Proc. of IEEE Int. Conference on Systems, Man & Cybernetics, Washington D.C., USA. 2003.

VIDEO PROC. IN INTERNATIONAL CONFERENCES

  • M. Prats, A. P. del Pobil, P. J. Sanz. A Control Architecture for Compliant Execution of Manipulation Tasks. Selected Video in Proc. of International Conference on Intelligent Robots and Systems, Be¼ing, China. October 9-15. 2006.
  • A. P. del Pobil, M. Prats, R. Ramos-Garijo, P. J. Sanz, E. Cervera. The UJI Librarian Robot: An Autonomous Service Application. Selected Video in Proc. of the IEEE Int.Conference on Robotics and Automation, Barcelona, Spain 2005.

INTERNATIONAL WORKSHOPS

  • A. P. del Pobil, M. Prats, P. Martinet, S. Lee, R. Dillmann. A Sensor-Based Approach for Dependable Physical Interaction. Technical Challenges for Dependable Robots in Human Environments. Pasadena, California, USA: 17-04-2008
  • M. Prats, P.J. Sanz, A.P del Pobil. A sensor-based approach for Physical Interaction based on Hand, Grasp and Task Frames. In Proc. of Robotics: Science and Systems – Robot Manipulation Workshop: Intelligence in Human Environments. Zurich, Switzerland, June 2008.
  • A. Morales, M. Prats, P. J. Sanz, A. P. del Pobil. An Experiment in the Use of Manipulation Primitives and Tactile Perception for Reactive Grasping. In  Proceedings of the Robotics: Science & Systems 2007 Manipulation Workshop - Sensing and Adapting to the Real World, 30 June, Atlanta, Georgia, USA, 2007.

Press

Software


Last modification by Webmaster
on 27/05/10

PROMETEO Research Group of Excellence sponsored by