I'm interested in every aspects concerning robot perception as the cornerstone of cognitive robotics.
Perception is the ultimate resource of an autonomous system, there is no autonomy no real faculty of reasoning, understanding and interpreting
the world without perception. On the other hand there is no understanding of what is perceived without the ability to represent, classify and
reasoning about the sensed world.
Visit our Laboratory
ALCOR
I'm currently working on:
1. The Gaze Machine: this is a wearable device that
identify the gaze scan path of the person wearing it, in
the environment. It can be worn quite easily, so the
subject wearing it can walk and go around in any light
conditions (I didn't say in the dark, it would be
meaningless) either indoor or outdoor.
Why it is important: we get lots of data that we are
using to model dynamic attention. We are working to make
these data available to the community.
2. Segmentation of human motion. There are harmonic
properties of human motion that makes it separable from
any other kind of motion.
3. Contrast sensitivity. 3D contrast sensitivity can sort
natural stimuli, in space and time. It might be used as
feature detector.
4. Interpretation of actions. This requires two main
steps: recognition and representation. Understanding
actions comes in a loop.
In the past I have been
working on theories of action and perception,
essentially with the Situation Calculus and on several aspects of reasoning like Abduction and Nonmonotonic reasoning.
|