Robotics, Brain and Cognition Laboratory
The Robotics, Brain, and Cognition Laboratory (RBCog-Lab) is a concentrated research infrastructure in the area of cognitive robotics. The core of the infrastructure consists of robotic platforms (hardware, software, and support crew) which includes one iCub robot, the most sophisticated humanoid robot in the world in terms of degrees of freedom. The ISR/IST host group is one of the developers of the original platform and hosts the only such robot existing in Portugal. RBCog-Lab includes additional robots (e.g. Vizzy, a wheeled robot with iCub-compatible software interfaces) and equipment such as motion capture systems and gaze trackers. The project was founded on a neuroscientific basis, keeping a strong link with this community during past and ongoing research projects.
Other recipients of our infrastructure come from a variety of fields such as robotics, engineering, developmental psychology, rehabilitation, education, and more.
We have participated in the development of the iCub robotic platform and we host the only instance of the iCub in Portugal. The iCub is an advanced humanoid robot equipped with 53 motors moving the head, arms, hands, waist, and legs. It includes cameras with human-like movement performance, stereo hearing capabilities, tactile and proprioceptive sensors. The iCub was designed to mimic a 3-year-old child. Its anthropomorphism is noticeable in its complex hands, facial expressions, and realistic movements of its waist, arms, and eyes. In addition to a full-body iCub, we also have two iCub heads mounted on fixed torsos (Chico and Chica).
Finally, the laboratory hosts a wheeled robotic platform (Vizzy) capable of tether-free movements, based on the same software platform as the iCub.
From left to right: full iCub robot, Chica, Chico, Vizzy.
Chica & Chico
The RBCog-Lab infrastructure features cutting-edge research tools to support 3D vision, video surveillance, robot calibration, navigation and manipulation, cognitive systems, and bio-inspired humanoid robots, including:
|Eye tracking system (Tobii)|
|Motion/gaze capture system (Optitrack)|
|Inertial Measurement Units (Xsens, LP Research)|
|3D sensors (Microsoft Kinect v1 and v2, Asus Xtion, Orbbec Astra, SoftKinetic DepthSense)|
|Cameras (Point Grey Research)|
|VR Headset (Oculus Rift, HTC Vive and Accessories)|
The robotic platforms of RBCog-Lab are used in several projects. In particular, due to the versatility and characteristics of the iCub, this robot was used to simulate human grasping, object interaction and manipulation in the real world (EU Projects First-MM, Handle), to study the way humans perceive the world (EU Project RoboSoM) and how they use natural language and visual action recognition as a learning mechanism to generate new behaviours and experiences (EU-Project POETICON++). For further information, please refer to our projects page.
The infrastructure is currently under the responsibility of the laboratory director (José Santos-Victor) and the permanent staff (Alexandre Bernardino and José Gaspar), who decide the relevant issues regarding the platforms, such as maintenance, upgrades and main research directions.
An engineering staff guarantees platform availability and conducts all necessary maintenance.
From left to right: José Santos-Victor, Alexandre Bernadino, José Gaspar
Our research group, together with our partners, is the main user of the infrastructure. These robots are sometimes used for showcasing and other events for the general public, but these possibilities are always considered under a cost/benefit analysis taking into account ongoing research constraints and deadlines at the time.
Access to the RBCog-Lab infrastructure is possible, pending
- A concrete experiment has been devised and implemented on the publicly available iCub Simulator (www.icub.org).
- The request is approved by the management staff. This will take into account the risk involved, possible lab resources that must be committed, the scientific merit of the proposal and the availability of the platform itself at the time.
This strict policy is required due to the non-trivial hardware involved which is able to self-destruct due to misuse. To minimize the consequences of a possible rejection, we recommend that you contact us first to obtain a pre-approval based on your proposal, but only projects implemented on the simulator which can be demonstrated to perform correctly will be granted access.
National RBCog-Lab network
A number of of well-known Portuguese research centres, active in the areas served by RBCog-Lab, have demonstrated interest in an active exploration of the RBCog-Lab as a research tool. This group of users will be managed in a dynamic manner and new members will be accepted once the access to the infrastructure is fully operational. The national institutional network is currently comprised of
- Institute for Systems and Robotics, Instituto Superior Técnico (ISR/IST), Universidade de Lisboa (host)
- Universidade de Coimbra, Institute for Systems and Robotics (Jorge Dias, Helder Araújo)
- Champalimaud Foundation (Zach Mainen, Adam Kampff, Rui Costa)
- Universidade de Aveiro (Luís Seabra Lopes, Victor Santos)
- Madeira Interactive Technologies Institute (Nuno Nunes, Sergi Bermúdez i Badia)
- Instituto de Medicina Molecular IMM (Tiago Maia, Leonel Luís)
- Faculdade de Engenharia, Universidade do Porto (A. Pedro Aguiar)
- Universidade do Minho (Estela Bicho, Wolfram Erlhagen)
Access to the RBCog-Lab infrastructure opens a door to collaborate with the whole cognitive robotics community, in particular with other institutions hosting similar platforms. The scientific goals proposed by IST are fully aligned with the future developments of the iCub platform and research community, namely with the Italian Institute of Technology (Professors Giulio Sandini, Giorgio Metta and Luciano Fadiga), the lead iCub Consortium partner. More info at www.icub.org.
RBCog-Lab has been selected to integrate the Portuguese roadmap of Research Infrastructures and is supported by the Fundação para a Ciência e a Tecnologia.