![]() ![]() |
![]() | ![]() ![]() ![]() |
Amy Loutfi, Silvia Coradeschi
In this work, we consider how to automatically create a correspondence between symbols, in this case linguistic names, and categories of odours using the data from an electronic nose. To accomplish this task, we create a system that first creates categories using unsupervised clustering and then generalizes from these catgeories by evaluating how well an odour name applies to its own sensor representations. Since part of the objective is to facilitate human and machine interaction, the names outputed from the system are correlated and with those of a human user. This may mean that in the context of an unsupervised categorization, perceptual differences between the human and the electronic nose may arise, however, these differences can be considered an asset especially when dealing with olfaction, for example when the electronic nose detects carbon monoxide that is normally perceived as odourless for a human. Therefore, to cope with the perceptual differences, the system is tuned to provide as explicit information about the categorization of odours as possible. We accomplish this task by adapting existing fuzzy-based algorithms to generate informative odour descriptions.
Keywords: Perception, Cognitive Modelling
Citation: Amy Loutfi, Silvia Coradeschi: Forming Odour Categories using an Electronic Nose. In R.López de Mántaras and L.Saitta (eds.): ECAI2004, Proceedings of the 16th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2004, pp.119-123.
![]() ![]() ![]() |