[robotics-worldwide] PhD Positions on Planning and Simulation of Grasping Under Uncertainty, Centro Piaggio, Univ of Pisa, Italy
m.gabiccini at ing.unipi.it
Sat Feb 16 02:49:02 PST 2013
University of Pisa, Italy
Research Center "E. Piaggio"
Two PhD positions are available on the following topics:
1. Active Haptic Information Gathering under Uncertainty + Grasp Planning Under Uncertainty
2. Advanced Simulation of Dexterous Manipulation Under Uncertainty
These activities cover some tasks within the
FP7-funded project PacMan ”Probabilistic and Compositional
Representations of Objects for Robotic Manipulation”. PacMan will
develop technologies to allow robots to generalize manipulation skills
to novel objects, and to cope with the uncertainty inherent in sensor
driven manipulation. This will be achieved by employing hierarchical,
compositional representations of objects (e.g. part based
representations of shape) that incorporate both visual and haptic
The appointees for the positions will work on the following topics:
1. Learning the Haptic Characteristics of Objects by in Hand Exploration (definition of control strategies for concurrent stable object grasping and object surface tracing / object surface rolling, frictional coefficient estimation, testing in simulation and on a real platform (KUKA LWR4+ and DLR-HIT Hand)) + Grasping Under Uncertainty (autonomous selection and execution of feed-forward pre-shaping hand-arm primitives within the soft-synergy paradigm, incipient grasp quality evaluation and local contact point refinement, grasp quality force optimization, grasp planning under uncertainty as a belief space planning problem)
A solid background in most of these subjects is required: efficient incremental representation of surfaces, machine learning and probabilistic methods of information propagation, grasp modeling, POMDP, strong coding (C/C++) and hardware/software integration skills.
2. Develop efficient and realistic numerical simulations for Grasping and Manipulation Under Uncertainty (incipient grasp quality evaluation and local contact point refinement, grasp quality force optimization)
Candidates for this topic should have an excellent background in: Rigid-body dynamics algorithms, Contact and Impact Modelling, nonlinear control, optimal control, experience with physics engines (ADAMS, Gazebo, OpenRave, Bullet, ODE), C/C++, parallel programming, GPGPU programming.
PhD studentship: full studentship available. Candidates should contact
us for details. You should have a master's degree in
computer science or mechanical engineering, together with
strong mathematical and coding skills. You should be in the top 10% of
your graduating class.
PacMan involves the following groups:
University of Pisa (www.centropiaggio.unipi.it)
University of Birmingham (www.cs.bham.ac.uk/go/irlab)
University of Innsbruck (iis.uibk.ac.at)
Applications (CV+letters of recommendation) should be made to Local Unit Coordinator Dr Marco Gabiccini (m.gabiccini at ing.unipi.it) and to the principal research supervisor
Prof. Antonio Bicchi (bicchi at centropiaggio.unipi.it). Please always copy Ms. Maley in email: (laura.maley at centropiaggio.unipi.it)
Please mark your email
"PacMan PhD Application"
Marco Gabiccini, PhD
Assistant Professor of Applied Mechanics and Robotics
Dept. of Mechanical, Nuclear and Production Engineering
University of Pisa
Via Diotisalvi 2, 56122 Pisa PI
More information about the robotics-worldwide