Welcome, Guest.
Please login or register.
Request Password
Student Projects and equipment
Student Projects
Students are invited to participate in several projects which are listed below.
Students are also welcome to propose other projects. To be persued these projects must be accepted by the BCBT2010 organization commitee.
1. Spatial attention for the iCub
Supervisors: Sergi Bermudez
Description: Generation of spatial attention using the idea of anticipatory gates for predictable dynamic objects is to be combined with temporal population coding for feature based object recognition. The idea would be to make the robot solve a shared attention task like the MakeNbreak game. As a first test, the usage of anticipations could be used to make the robot predict the endpoint of a moving ball and stop it by moving the hand towards that point on the flat plane (table).
Students : Andre Luvizotto, Zenon Mathews, Giacomo Spigler, Marco Antonelli
2. Precise motor control using a cerebellar model
Supervisors: Ulysses Bernardet, Marti Sanchez
Description: Cerebellar research is often related to motor control, but realistic cerebellar models have seldom been applied to robotics. Robots that accomplish complex tasks usually employ controllers only vaguely inspired in the cerebellum, and more realistic models are used only to solve simpler problems as eyeblink conditioning. The project is to extend an abstract version of an existing cerebellar model in order to perform a control task with a only a few degrees of freedom (1 or 2) and an analog error signal, such as the vestibulo ocular reflex or smooth pursuit. The ultimate goal is to run the model on the iCub robot.
Students: Ivan Herrero, Andrew Martin, Antoine Deleforge, Damien Drix, Friederix Johansson, Jennifer Lewis, Tomas Voegtlin.
3. MiX-XRA: decision making (DAC)
Supervisor: Marti Sanchez
Description:
Students:
4. BCI control of the Robot (RoboNova)
Supervisor: Sergi Bermudez
Description: The brain reacts to visual stimulation that flickers at specific frequencies, generating neural activity in the visual cortex at the same frequency of that of the stimulus. This responses are known as Steady State Visual Evoked Potentials (SSVEP). Using Electroencephalography (EEG) it is possible to read brain's activity and detect SSVEPs. The main idea of this project is to make use of the SSVEPs to control the movement of a biped robot (Robonova-1) in real time, connecting the human brain to a robot.
Students: Arnau Espinosa
5. Robot navigation: tankDZR
Supervisor: Sergi Bermudez, Marti Sanchez
Description:
Students: Cesar Renno-Costa, Encarni Marcos, Melle Hofman, Gerd Gruenert, Lorena Perez Mendez
6. Real-time visualization of neuronal system in the eXperience Induction Machine (XIM)
Supervisor: Ulysses Bernardet
Description: The main is to use XIM as a visualization and navigation platform for understanding the structure and dynamics of neuronal systems. One option is that there is a 3D representation of an iqr system, that the user can navigate in real-time. Depending on the position within the system, information about the system is displayed, e.g. time or group plots.
Students: Alberto Betella, Rodrigo Guedes-Carvalho, Nicole Lazzeri
7. Neurobiologically inspired real-time interaction with video portraits in the eXperience Induction Machine (XIM)
Supervisor: Anna Mura, Ulysses Bernardet
Description: The eXperience Induction Machine (XIM) is a multi-user mixed-reality space at the SPECS lab (click here for more infos). Goal of this project is to develop a neurobiologically inspired system where a user in XIM interacts with video portraits in real time The system will build up on the existing infrastructure, and interactive systems. The neuronal system logic will be implemented with the neuronal systems simulator iqr.
Students: Henrique Serro, Daniele Mazzei
8. Visual search in images with emotional content
Supervisor: Sergi Bermudez
Description:
Students: Cristina Campillo, Jose Maria Blanco, Xerxes Arsiwalla
9. Build smart camera
Supervisor: Gerd Gruenert (gerd.gruenert [at] uni-jena [dot] de
Description: compose final image out of (smaller/low resolution) images...
Students:
10. Closed-loop multi-modal stimulus generation for RGS?
Supervisor: Sergi Bermudez
Description:
Students: Sylvain Le Groux, Monica Camerairo, Belen Rubio
11. Construct a cortex-like Recurrent Neural Network with Evolutionary Algorithms
Supervisor: 
Xavier Hinautxavier.hinaut [at] inserm [dot] fr">
Description: Cortex is highly structured, organised in layers (laminar structure) and has different connections pattern for each layer. The goal of this project is to try to understand why this structure is efficient and how information is processed in such networks. We will use Evolutionary methods like Genetic Algorithms to generate networks that have a cortex-like topology, to see how it improves the performance of sequence discrimination and categorization tasks.
Reservoir Computing (see http://www.scholarpedia.org/
Students: Xavier Hinau, Man Yi Yimxavier.hinaut [at] inserm [dot] fr">
12. Control of XIM via BCI
Supervisor: Hannu Jarvinen (hannu.jarvinen [at] tml [dot] hut [dot] fi)
Description:
Students:
Equipment
Students should request to Ulysses Bernardet and Sergi Bermúdez the equipment needed for their projects.
- iCub robot
- e-puck robots
- ground robots
- 20+ cameras
- wii remote controllers
- Physiological recording systems
- Tobii 120X eye-tracker
- Mobilab with sensors for:
respiration
ECG (cardiogram)
facial EMG (miogram)
pulse
EDA (electrodermal activity) - Software:
iqr neural simulator
Matlab
Simulink
E-prime (+eye tracking)
Presentation
