Biologically inspired feature extraction, sensory information fusion and perception methods

Measuring has been, without a doubt, a very old need for human beings. However the concept of the multi measuring i.e. artificial multi sensing is very new and this has appeared almost together with the concept of artificial perception and the discussion on biological inspiration. The main goal of this project is to develop and apply methods for feature extraction and systems for sensory information fusion, artificial perception to be used in industrial applications oriented especially to the food industry and intelligent rescue systems (IRS). Until now the traditional analysis techniques (PCA, median filter, etc.) has been mainly used for feature extraction requirements and a number of combination of artificial neural networks and fuzzy logic based models have been applied for the fusion of the sensory information originating from two (taste-smell and auditory-visual) sensory devices. In the current state of the project a hybrid ANN structure is being tested to handle the sensory outputs of the feature extraction phase based on wavelet transform techniques. Moreover, a study on a new multi sensor fusion model including the active perception is being conducted.

	Selim Eskiizmirliler
	Linn Robertsson
	Iasen Hristozov
	Peter Wide

Assi Domän - Frövi

	Biel L. and Wide P., "Active Perception for Autonomous Sensor Systems", IEEE Instrumentation & Measurement Magazine, December 2000, pp. 27-30.
	Wide, P. and Winquist, F., Bergsten, P. and Petriu, E. M, "Human-based multi-sensor fusion method for artificial nose and tongue sensor data.", IEEE Transaction of Instrumentation & Measurement Technology, vol. 47, No. 5, Oct. 1998, pp. 1072-1077.