ALCOR

                                                                            Autonomous Agent Laboratory of cognitive Robotics                                        12/08/08

Home
Research
Publications
Teaching
ALCOR
Software

 

Alcor Laboratory is located at Quarto Miglio (fourth mile of the Appia Antica), via San Tarcisio 77.
Here is the web page (still under construction) of the ALCOR research group

Here are some pictures of the laboratory  with built in rescue arenas mosaic, lab1, lab2, some of the mannequins we use for victim recognition in rescue arenas.

People

Projects:

          LISBON 2004Text Box: Visiting the laboratory.
Students from other countries that want  to visit the laboratory are very well come, if their curricula is consistent with the current projects. We can support them just for the trip and help them to find accomodations, we cannot offer any other kind of support.

          OSAKA 2005

  • ENEA: RAS (Surface autonomous robot)

          Ras1, Ras2, Report-2003

  • Galileo: Flir-OWS


HISTORY

The ALCOR laboratory was founded in 1998 at DIS
(Dipartimento di Informatica e Sistemistica, Università di Roma “La Sapienza”) with the aim at
understanding the rational of  a cognitive robotics system, building a full autonomous system.

Starting with simple tiny robots we have been developing, over more sophisticated platforms, a series of autonomous systems.
Armandino1, Armandino2, Armandino3
Mr. ArmandX
DORO
Shrimp


ALCOR mission is now to develop autonomous systems capable of an intelligent interaction with the environment.
A system showing an intelligent interaction is one manifesting awareness of the environment  and being attentive and driven by curiosity and interest for its surroundings, recognizing places, some class of objects, and akin to learn to manage its behaviours.

 We mainly focus on endowing an autonomous system with a sophisticated sensory apparatus,  and with an adequate
data modeling and elaboration structure deploying perception, reasoning with time-uncertainty-modalities, learning, and decision theoretic planning, in order to execute complex tasks.


The scientific methodologies adopted, from the Artificial Intelligence Community, serve to develop the following autonomous skills:
Attentive  Perception: stereo-cameras on a pan-tilt head, coupled with lasers on their own pan-tilt, and thermocamera.
Proximity and guidance:  sonar, inertial platform, laser range finders,  odometers.
Active and dynamic learning for objects recognition, for learning concepts and behaviours, using classification, reward and imitation based techniques.
Decision-theoretic methods utility based to determine behavioural strategies.
Mission planning and scheduling for goals and resources management and
Reactive planning to control short-range behaviours.
Cognitive architecture to give the system a full range power of integration and a high level capability to manage its own different functionalities.

Mobility

Presently only the shrimp (platform from bluebotics) is adequate for mobility on rough terrains.


 

 

Home | Research | Publications | Teaching | ALCOR | Software

This site was last updated 01/09/06