Research topics

This page enumerates the research topics that I am actively involved with. If you have interests on any of these areas, please contact me.

Cognitive architectures

In the engineering of machine intelligence one cannot disregard natural inteligence that humans, together with other animals, display. Intelligence is here understood in the broad sense of behaving appropriately situated in an environment, as in the case of a fight-or-flight response for securing survival, as well as in the case of a mathematician proving a theorem. In these situations, hard to cast as a neat optimization problem, scientists often look into biology for inspiration. What biological evolution solved before has the potential for, at least, provide guidance on how to approach the problem.

My interests in this area are focused on (1) decision-making, namely in the interplay between emotion mechanisms and rationality, and on (2) robotic platforms (e.g., humanoid robots) situated in a physical environment, socially interacting with humans, other robots, and objects in the environment.

Field robotics

Unstructured environments pose a challenging problem for mobile robots, as traditional approaches that perform wellon flat, horizontal ground with well defined obstacles, can no longer be directly applied. Moreover, being able to address for this kind of scenarios extend the field of application of robotics to areas with major social impact, such as response to a natural (or artificial) disaster in a search and rescue (SAR) scenario. In these scenarios it is paramount to be able to find possible victims in, for instance, a building near to collapse.

I am currently leading a research team focused on a robotic platform designed for SAR scenarios: the RAPOSA robot. This robot is the outcome of a successful collaboration between a Portuguese company and the university. RAPOSA was developed in a consortium led by the company IdMind, in partnership with Instituto Superior Técnico (IST), which currently owns the platform.

Current work includes automatic stair climbing, and a HMD user interface.

Human-robot interaction

Negleted for a long time, the interaction between humans and robots is becoming an hot topic on robotics. This broad research area, encompassing social interaction among robots and humans, as well as efficient teleoperation of robots, pose many challenges to the community.

In this area I am particularly interested on endowing SAR robots operators with the interface capable of providing the best situation awareness possible. This includes aspecs such as immersiveness, effective data display, and adjustable autonomy.

Automatic karyotyping

One of the medical tests to diagnose leukemia requires a cytogenetics technique called karyotyping. This technique consists on obtaining an microscope image of cell division, in such a way that the chromosomes become visible. The diagnostics of leukemia, for instance, involve visual inspection of the chromosomes by medical staff, in the search for anomalies. This inspection is extremely time consuming, as the technician has to manually group (using a GUI and proper software) all chromosomes according to their classes.

I have been involved on a research effort to develop automatic techniques for this grouping operation. The automatization of this process brings significant benefits because of the heavy demand on this test and the time needed to perform it. Advanced biomedical image processing techniques have been used in this research, together with numerical optimization algoritms.

Automatic music transcription and search

My passion for music has always pushed me towards the symbiosis of technology and music. I have been involved on the research of techniques to, given a human whistle of a musical theme, transcribe it to a MIDI representation, and search for similar themes on a database of MIDI files, ranking the candidates according to a similarity metric (relevance). Techniques used to accomplish this task include: robust pitch detection, note detection, string encoding of notes, and string similarity measures.