printlogo
http://www.ethz.ch/index_EN
Welcome
 
print
  

Marques, Hugo Gravato, Dr.

ETH Zürich
Dr. Hugo Gravato Marques
Institut für Robotik und Intelligente Systeme
TAN C 4
Leonhardstrasse 27
8092 Zürich
SCHWEIZ

Phone: +41 44 632 43 24
Fax: +41 44 632 14 68
E-Mail: 

Interests

My current interests lie on the kind of self-organization processes that can give rise to coordinated motor activity and appropriate decision making. More specifically i'm interested in understanding general self-organization principles in mammals, with especial focus on the sensorimotor development of the spinal cord, and how these principles can contribute to the construction of more advanced (open-ended) robotic systems.

Short CV

Education

Professional experience

Publications

Journal

Conferences

Students

PhD (co-supervision): Cristiano Alessandro, Naveen Kuppuswamy

Master: Simon GAY, Farhan Imitiaz, Konstantin Zerebcov, Igor Sokolovski, Christian Hückstädt, Stefan König, Kristin Völk, Stefano Toxiri, Arjun Bharadwaj, Christoph Bolliger, Michael Ackermann, Tobias Klauser

Bachelor: Christian Brändli, Fabian Koeth, Mathias Weyland, Philip Shaffner, Aditya Pralhad Patil, Cordt Voigt, Mathias Wild, Imanol Studer

Robots (under construction)

roboy

Roboy: A Manufactured Musculoskeletal Robot (2013)

History: Roboy is a musculoskeletal robot built using manufacturing techniques. The robot was the outcome of a colaboration between the AI Lab and several Swiss companies.
Research: Like the EDS, Roboy is mainly a show robot, but it was also intended to serve as an initial prototype of a stable research platform. I particilpated in the mechanical design as well as in programming the robot.
Publications: Pfeifer, Marques and Iida 2013
Details: video

 
DVSTrolley

DVS Car (2012)

History: The DVS Car is equipped with an inertia measurement unit (IMU) and an assynchronous camera which outputs spiking data in a way analogous to the human retina (DVS). The camera is mounted on a pan and tilt unit. The goal of this robot was to investigate the self-organization of optic-flow columns and the self-organization of the vistibulo-ocular reflex (VOR).
Research: I built this robot and together with Fabian Koeth, showed that the topological organization of the optical-flow columns, similar to those observed in the MT area of the brain, can be self-organized.
Publications: Koeth and Marques, 2013
Details: video

 
hand

Skinned Hand (2012)

History: The robotic hand (name) was equipped with a composed of 24 tactile sensors to investigate the self-organization of somatosensory topological information.
Research: Together with Konstantin Zherebtsov, I showed that ON-OFF center-surround receptive fields (similar to those in the visual cortex) could be obtained using simple correlations between the tactile signals.
Publications: not published
Details: video

 
pendulum

Tendulum (2011)

History: The Tendulum robot was built by me and Cristiano Alessandro to serve as a simple platform to investigate autonomous motor development in musculoskeletal robots.
Research: Together with Mathias Weyland, I used this robot to self organize analogues to four spinal reflexes: the myotatic, the reciprocal inhibition, the reverse myotatic, and the withdrawal reflexes. These reflexes are carried out in three sensor modalities: muscle length, muscle force and cutaneous.
Publications: in preparation
Details: video

 
eds

EDS: The ECCEROBOT Design Study (2010)

History: The EDS, also built by Rob Knight, was the successor of CRONOS. The EDS has all the low-level processing inside the robot, which avoids a large number of cables from coming out of the robot.
Research: This robot was more of a show robot. I helped programming and accompained it to several events in Europe.
Details: not published
Details: video

 
cronos

CRONOS: A Musculoskeletal Robot (2007)

History: CRONOS is an anthropomimetic robot built by Rob Knight. The goal of the robot is to mimmic as close as possble the human musculoskeletal system.
Research: Together with Owen Holland, I used the robot during my PhD to investigate the interaction between the robot and a physics simulation of the robot in the context functional embodied imagination.
Publications: Marques et al. 2010
Details: video

 

Simulations (under construction)

nmssimulator

NMS Simulator Library (2012)

History: The neuro-musculo-skeletal (NMS) simulator library is a group of Simulink and SimMechanics blocs that can be easily assembled to simulate musculoskeletal systems.
Research: I built this library from scratch. Together with Arjun Bharadwaj, I'm working on a musculoskeletal leg model which uses self-organized reflexes to coordinate the muscle activity during hopping.
Publications: in preparation
Details: video

 
caliper

Caliper (2011)

History: Caliper is a simulator mainly built by Steffen Wittmeier, Michael Jäntsch and Konstantinos Dalamagkidis. The simulator, which is based on Bullet physics, was built as a general platform to simulate different types of tendon driven systems, both natural and artificial.
Research: Together with Farhan Imitiaz, I used this simulator to self-organize analogues of three spinal reflexes: the myotatic, the reciprical inhibition, and the reverse myotatic reflex.
Publications: Wittmeier et al 2011; Marques, et al 2013
Details: video

 
simnos

SIMNOS (2007)

History: SIMNOS was built by Richard Newcombe to simulate the mechanical interactions in CRONOS.
Research: Together with Richard Newcombe and Owen Holland, I used the robot during my PhD to investigate simple controllers in for musculoskeletal systems as well as to investigate the interaction between the CRONOS robot and its internal model (SIMNOS) in the context functional embodied imagination.
Publications: Marques et al 2007; Marques et al 2008
Details: video

 

Videos (to be moved)

Reflex Learning and Jumping in Simulated Leg Model (2012)

In this video we use a simulated leg model with 6 muscles to test our reflex learning framework. The framework is capable of identifying automatically the connectivity of the Myotatic and Inhibition reflexes. Once the reflexes have been learned, the sensory stimulation induced by the falling leg is enough to make the leg jump.
References: Marques, H.G., Völk, K., König, S., and Iida, F. (2012). Self-organization of spinal reflexes involving homonymous, antagonist and synergistic interactions. In Proceedings of SAB 2012, 27-31 August, Odense, Denmark.

Reflex Learning in Tendon-Driven Robot (2011)

In this video we use agonist-antagonist pair of artificial muscles to test our reflex learning framework in a tendon-driven peundulum robot. With our framework we have self-organized 4 spinal reflexes -- the Myotatic, the Reciprocal Inhibition, the Reverse Myotatic and the Withdrawal reflex -- in three different sensor modalities: muscle length, muscle force and tactile. The pendulum is immersed in water to simulate the viscuosity of uterus environment.
The video shows first the process of single muscle twitches used for learning the reflexes, and after the testing of the each of the reflexes:
(1) The Myotatic and Reciprocal Inhibition reflexes are tested by stretching one of the muscles; the length increase in one muscle causes the muscle to contract and the antagonist to relax.
(2) The Reverse Myotatic reflex is tested by setting a constant muscle activity to which the activity produced by the reflex circuit is added. The increase in muscle tension reduces the motor activity and prevents the muscle from reaching excessive forces.
(3) The Withdrawal reflex is tested by poking one of the tactile sensors. The tactile activity produced in one tactile sensor leads the contralateral muscle to contract and inhibits the ipsilateral muscle. This activity results in a withdrawal of the limb from the contact area.
References:: in preparation

 

Wichtiger Hinweis:
Diese Website wird in älteren Versionen von Netscape ohne graphische Elemente dargestellt. Die Funktionalität der Website ist aber trotzdem gewährleistet. Wenn Sie diese Website regelmässig benutzen, empfehlen wir Ihnen, auf Ihrem Computer einen aktuellen Browser zu installieren. Weitere Informationen finden Sie auf
folgender Seite.

Important Note:
The content in this site is accessible to any browser or Internet device, however, some graphics will display correctly only in the newer versions of Netscape. To get the most out of our site we suggest you upgrade to a newer browser.
More information

© 2013 ETH Zurich | Imprint | Disclaimer | 1 May 2013
top