| 04-10-25 |
Puri Cariñena Amigo
(Intelligent Systems Group
Department of Electronics and Computer Science
University of Santiago ) |
Model of Fuzzy Temporal Rules
Abstract: Generally fuzzy systems (particularly in control/monitoring environments)
are based on fuzzy if-then rules which do not include time as an explicit
variable.
There exist aspects related to the semantic richness of the expressions used
by a human expert which are not contemplated by these models. Generalization
of fuzzy rules to Fuzzy Temporal Rules (FTRs) allows the representation and
explicit handling of time, facilitating the design of solutions for problems
associated to dynamic systems and processes.
I will briefly describe in this talk the grammar we have developed for the
propositions making up the FTRs. The proposed model contemplates crisp and
fuzzy variables, with absolute and relative (fuzzy or not) temporal and
value constraints on them, quantifiers (allowing to model the concept of
persistence of a value in an interval), and operators (last, minimum, mean,
∑) on the variables.
This model of FTRs has already been used in real applications in mobile
robotics, in tasks such as wall-following, mobile-obstacle avoidance and
landmark detection. Our next objective is to apply this paradigm in the
field of medicine, in particular to the task of automatic knowledge
extraction, using data from a cardiology unit. |
| 05-02-28 |
Simon Case (Senior Research Scientist,
Computational Intelligence Group, IS Lab
BT Research & Venturing) | Agent Mediated Knowledge Management in a Telecommunications Service Provision and Maintenance Domain |
| 05-03-16 (16:00 - 16:50) |
Murray Shanahan (Reader in Computational Intelligence,
Dept. Electrical & Electronic Engineering,
Imperial College London) |
A Brain-Inspired Architecture for Cognitive Robotics |
