Επιτεύγματα

Artificial systems with perceptual capabilities and robotic systems in general are complex combinations of mechanical and electronic parts that operate through the coordinated activation of perceptual capabilities, control strategies and learning techniques. All of these components have to be appropriately integrated in a common framework. The effectiveness of each of the subsystems cannot be fully assessed in isolation. This is because the success of the overall system does not depend on the quality of the constituent elements only, but also on their appropriate integration and on the characteristics of the interaction of the resulting system with the environment. Integration in this context is an important problem per se since it has to cope with considerable complexity and, in many cases, with scarce resources and conflicting requirements. For this reason, a lot of effort has been devoted to the development of integrated prototype systems and demonstrators. The importance of such prototypes is twofold; Firstly, they act as test-beds for evaluating the true effectiveness and robustness of their constituent components. Secondly, each of these prototypes addresses the needs of important application areas. Four such prototypes that have been achieved at CVRL are briefly described below. These research prototypes have been successfully tested and validated through extensive user trials at a national but also international level. Moreover, they have attracted the interest of the general public, through the extensive coverage they got in the national and international press and other media.

  • Στο παρακάτω άρθρο απονεμήθηκε το "Best paper award" συνοδευόμενο από και από χρηματικό έπαθλο US $500.
    A.A. Argyros, M.I.A. Lourakis, "Vision-based Interpretation of Hand Gestures for Remote Control of a Computer Mouse", in proceedings of the HCI 06 workshop (in conjunction with ECCV 06), LNCS 3979, Springer Verlag, pp.40-51, Graz, Austria, May 13th, 2006.
  • TOURBOT, an advanced robotic tour-guide used in museums and other exhibition sites. TOURBOT is able to autonomously map the environment and then offer guided tours to on-site and web visitors. TOURBOT is capable of operating robustly in crowded environments and communicates with on-site visitors through computer-generated voice messages, as well as through facial expressions that externalise its internal state.
  • A prototype skin-colour tracking system that performs real-time 3D tracking of multiple skin-coloured regions in the field of view of a potential moving observer. This system has been developed and used in the context of the EU-IST project ActIPret where the goal was to develop a cognitive vision system able to automatically interpret the activities of people handling tools. Moreover, the same tracking algorithms have been used to develop a stand-alone human-computer interaction system that permits to a human to control an MS-Windows computer through hand gestures.
  • A camera tracking system that performs on-line estimation of 3D camera position and pose in parallel with 3D scene reconstruction.
  • SBA, a generic sparse bundle adjustment software package (http://www.ics.forth.gr/~lourakis/sba). Bundle adjustment amounts to a large-scale minimization problem that is almost invariably used as the last step of every feature-based structure and motion estimation computer vision algorithm to obtain optimal 3D structure and viewing parameter estimates. Due to the large number of unknowns contributing to the minimized objective function, general-purpose numerical optimization codes incur high computational costs and are very demanding in terms of memory storage when applied to the problem of bundle adjustment. Fortunately, the lack of interaction among parameters for different 3D points and cameras in multiple view reconstruction results in the underlying normal equations exhibiting a sparse block structure, which can be exploited to gain considerable computational benefits. SBA is a publicly available C/C++ software package for generic bundle adjustment employing a sparse variant of the Levenberg-Marquardt algorithm. The package can be very useful to researchers working in fields such as computer vision, robotics, image-based graphics, photogrammetry, surveying, etc. To our knowledge, it is the first and currently the only such software package to be offered free of charge.
  • SIMUUN, a simulation environment for undulatory locomotion studies in robotics and biology (http://www.ics.forth.gr/~tsakiris). SIMUUN is a block-based simulation environment based on the Matlab/Simulink software suite. It has been developed and is being used in the context of the EU IST-FET project BIOLOCH, which is related to the study and implementation of novel biomimetic robotic locomotors as tools for medical robotics (in particular, for locomotion in the tortuous environment of the human gastrointestinal tract).
  • DRIVER, a robotic wheelchair for people with disabilities, possessing semi-autonomous navigational capabilities. DRIVER can perform collision avoidance, autonomous navigation in corridor environments and person following at a fixed distance. DRIVER has an advanced multimedia man-machine interface that facilitates the control of the overall system. DRIVER also obeys to voice commands, provided that the execution of such commands does not pose a threat to the user.
  • Best Presentation Award from the Organization of Computational Neurosciences.
    D.P. Tsakiris, A. Menciassi, M. Sfakiotakis, G. LaSpina, and P. Dario, ''Undulatory locomotion of polychaete annelids: mechanics, neural control and robotic prototypes''. Extended abstract, poster and presentation at the Annual Computational Neuroscience Meeting (CNS*2004), Baltimore, USA, July 17-22, 2004.
Όροι Χρήσης | Πολιτική Προστασίας Δεδομένων Προσωπικού Χαρακτήρα