1. Presentation
|
|
SAM (since Oct. 2010)
|
SAM (before Oct. 2010)
|
SAM (Smart Autonomous Majordomo) is a mobile robot with an arm and a
gripper. It is composed of a ROBULAB 10 mobile base from the company ROBOSOFT, on which is mounted a 6-DOF MANUS arm from the company EXACT DYNAMICS which ends with a gripper.
SAM is used for assisting disabled/ageing persons in their environment.
It can fetch objects and bring them back to a disabled person. It can
also propose physical and/or cognitive exercises to him.
2. Members
Group leader: Christophe Leroux (LRI).
Members (as of April 2011) : Patrick Hède (LVIC), Philippe Morignot (LRI), Mariette Soury (LRI), Mohamed Ben Ghezela (LRI).
Past members: Héléna Vorobieva (LISA), Anthony Rémazeilles (LRI), Nathalie Lot (LRI), Anthony Truchet (LRI).
3. Projects
ANSO (05-07), AVISO (07-09), MIDAS (09-11) for ITEA 2, ARMEN (10-13) for the European Commision..
4. Technical Features
More technical explanations can be found in our scientific publications (available here, here or here).
4.1. Visual Servoing
Anthony Rémazeilles. Under construction, sorry...
4.2. Object recognition
Patrick Hède works on PIRIA,
an object recognition library, which can efficiently recognize the
object image which is the closest to another object image from a
database.
4.3. Object manipulation
Héléna Vorobieva worked on determining the pose for grasping an object
by its smallest surface. For example, a soda can can be grasped from
any direction. A card deck or a mug must be grasped under a specific
angle.
4.4. Planning
Philippe Morignot works on scenario generation and execution for SAM.
Goals are specified by a user (e.g., a tetraplegic person) and an
initial state of the environment is supposed to be known (together,
they constitute a planning problem, expressed in PDDL, see a
pick-and-place example here). SAM's possible actions are expressed as PDDL operators (see an example of domain here). Then the Constraint Programming Temporal task planner (version 2, written by Vincent Vidal (ONERA Toulouse)) hopefully generates a plan (see an example here). This plan is turned into a scenario by attaching handlers (see an example here) to the plan's instantiated operators. This scenario is expressed as an XML file (see an example here).
This scenario is interpreted by an Interactive Scenarization Engine
(ISEN), written by Christophe Leroy (CEA LIST, LSI), which produces
behaviors of the robot SAM (see a video here or there).
Last edited: April 2011.
pmorignot@yahoo.fr
