[Amdnl] PhD positions avialable at the Italian Istitute of Technology to work on the iCub

[Giorgio Metta]

Dear roboticists,
This is to advertise the following Phd positions with scholarship at the iCub Facility, Istituto Italiano di Tecnologia.

The positions are available through the PhD course of Bioengineering and Robotics, curriculum on Advanced and Humanoid Robotics. Successful applicants will join the iCub Facility department at the Istituto Italiano di Tecnologia in Genoa (http://www.iit.it/icub). For additional details please check: http://www.iit.it/en/openings/phd-calls.html

********
Theme 1: Articulatory information for automatic speech recognition in human-robot verbal interaction
Tutors: Giorgio Metta, Leonardo Badino

Description: 
State-of-the art Automatic Speech Recognition (ASR) systems produce remarkable results in some scenarios but still lags behind human level performance in several real usage scenarios and perform poorly whenever the type of acoustic noise, the speaker's accent and speaking style are "unknown" to the system, i.e., are not sufficiently covered in the data used to train the ASR system. The goal of the present theme is to create a robust speaker-independent Key Phrases Recognition system where commands delivered by a user to a robot (i.e., the key phrases) have to be recognized in unconstrained utterances (i.e., utterances with hesitations, disfluencies, additional out-of-task words, etc.), in the challenging conditions of human-robot verbal interaction where speech is typically distant (to the robot) and noisy. To improve the robustness of the recognition system to intrinsically challenging and/or unknown conditions we will create recognition systems that combine acoustic features with predicted vocal tract articulatory movements (Canevari et al. 2013) thus exploiting the articulatory representation of speech where phonetic targets (i.e., the articulatory targets necessary to produce a given sound) are largely invariant and complex phenomena observed in speech are can be compactly modeled.

Requirements: 
The successful candidate will have a degree in computer science, bioengineering, physics or related disciplines. A background in speech and language processing is a plus.

Reference: 
[1] Canevari, C., Badino, L., Fadiga, L., Metta, G., "Cross-corpus and cross-linguistic evaluation of a speaker-dependent DNN-HMM ASR system using EMA data", Interspeech Workshop on Speech Production for Automatic Speech Recognition, Lyon, France, 2013.

Contacts: leonardo.badino at iit.it, giorgio.metta at iit.it

********
Theme 2: Whole-body real-time motion planning and control in compliant humanoids: theory and implementation on the iCub
Tutors: Francesco Nori, Giorgio Metta, Daniele Pucci

Description: 
the iCub is among the few humanoids capable of whole-body motion control while being compliant to perturbations (https://www.youtube.com/watch?v=VrPBSSQEr3A). These capabilities are possible thanks to the iCub peculiar sensors such as whole-body distributed tactile and force sensing. These sensors allow the iCub to simultaneously control joint movements, joint torques and interaction forces. Within this context, motion control is still an open research activity since previous approaches mainly dealt with kinematic movement planning. The iCub instead needs a dynamic planner to simultaneously plan movements, torques and interaction forces. The goal of the proposed theme is to develop a kinematic and dynamic motion control software. Given a control task, the foreseen controller will compute not only the kinematics (i.e. joint trajectories) but also the dynamics (i.e. joint torques and interaction forces) to perform the task. Among the various solutions to the problem, optimal control will represent a primary focus of attention. Specific tasks will include multi-contact motion planning [2], whole-body motion planning [3] and non-rigid contact motion planning. The project will be conducted within the context of a European project and the candidate will be asked to spend a six months experience in a foreigner country as a visiting Ph.D. student.

Requirements: 
candidate is requested to have a background in control theory, computer science and/or robotics. Specific background in optimal control will be positively evaluated.
Special requirements: this theme is supported by an ITN-H2020 research grant (SECURE). Candidates should satisfy the eligibility criteria of the Marie Skłodowska-Curie Actions. The position is suitable for candidates who have not have resided or carried out their main activity in Italy for more than 12 months in the last 3 years preceding their recruitment.
Note: the position is associated with a considerably higher salary than a regular studentship and offers many opportunities to travel to the other European partners in the network.

Reference:
[1] F Nori, S Traversaro, E Jorhabib, F Romano and D Del Prete A.; Pucci. iCub Whole body Control through Force Regulation on Rigid Noncoplanar Contacts. Frontiers in Robotics and AI.
[2] A Escande, A Kheddar, S Miossec. Planning support contact-points for humanoid robots and experiments on HRP-2. 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems.
[3] Moritz Diehl, Hans Joachim Ferreau, and Niels Haverbeke. Efficient Numerical Methods for Nonlinear MPC and Moving Horizon Estimation. Nonlinear Model Predictive Control. Lecture Notes in Control and Information Sciences Volume

Contacts: francesco.nori at iit.it, giorgio.metta at iit.it, daniele.pucci at iit.it

********
Theme 3: Body representations learning for whole body control
Tutors: Giorgio Metta, Francesco Nori

Description: 
The goal of this PhD theme is to develop complex body representations for the iCub robot utilizing visual, tactile, force, and motor signals. This representation enables control of the robot in the presence of multiple potential or actual contacts. It will e.g. determine that certain configurations are convenient to support the robot during task execution and actively seek contact with the environment rather than avoiding it. This model will be also compared with an existing neuroscience model of the so-called peripersonal space representation. The main result will be a set of software libraries that contain learning methods (e.g. probability density estimation, regression, etc.) and sensory processing (e.g. optical flow estimation, stereo vision, etc.). These algorithms will jointly contribute to the acquisition of the body and peripersonal space representation. Specifically for this theme, we will seek to build controllers for the whole body of the robot (legs, arms, etc.) which use the acquired sensorimotor representations.

Requirements: 
candidate is requested to have a background in control theory, computer science and/or robotics. Specific background in optimal control will be positively evaluated.
Special requirements: this theme is supported by an ITN-H2020 research grant (SECURE). Candidates should satisfy the eligibility criteria of the Marie Skłodowska-Curie Actions. The position is suitable for candidates who have not have resided or carried out their main activity in Italy for more than 12 months in the last 3 years preceding their recruitment. Note: the position is associated with a considerably higher salary than a regular studentship and offers many opportunities to travel to the other European partners in the network.

Reference:
[1] Roncone, A.; Hoffmann, M.; Pattacini, U.; Metta, G. Automatic kinematic chain calibration using artificial skin: self-touch in the iCub humanoid robot. IEEE International Conference on Robotics and Automation (ICRA2014) pp.2305-2312 Hong Kong, China, May 31-June 7, 2014
Contacts: giorgio.metta at iit.it, francesco.nori at iit.it

********
* How to apply

Please note that the positions are available through the PhD course of Bioengineering and Robotics, curriculum on Advanced and Humanoid Robotics, offered jointly by IIT and the University of Genoa. The official call and application forms are available on the website of the University of Genoa.

The official calls are available here:
http://phd.dibris.unige.it/biorob/index.php/how-to-apply/themes (Advanced and Humanoid Robotics)

Applications must be submitted online, instructions for applicants are available here:
http://phd.dibris.unige.it/biorob/index.php/how-to-apply

* Application deadline: 10 June 2015, 12pm Italian time

* Applicants are strongly encouraged to get in touch the contact person(s) for the individual themes.

********
Regards,

Giorgio Metta <giorgio.metta at iit.it>

Istituto Italiano di Tecnologia
Via Morego, 30                            
16163 Genoa, Italy                         
Ph: +39 010 7178-1411
Fax: +39 010 7178-1205

URL: http://pasa.liralab.it