Advanced Recognition System

01.11.2004 - 31.10.2005
Auftragsforschungsprojekt
For over 100 years neurologists, psychologists and pedagogics (NPP) have made huge progress in the scientific study of the human brain. Especially the last 20 years have brought amazing results to light, which should encourage applying them in the bionic field as well. We are convinced that this new way will help to restore Artificial Intelligence (AI) in many fields. From the current view of NPP it is obvious, that consciousness and human consciousness are not describable by classical mathematical algorithms. Where are concrete applications? We need methods to design and to integrate sensor and actuator networks with hundred thousands of nodes. We have to maintain them in a cost-efficient way. We have to develop powerful, highly flexible control systems, to give answers for future demands in the automation area. In this sense we address themes like energy saving, geriatric health care, facility management, efficiency improvement in hospitals, and many others. The decisive point is, that complex scenarios must be compiled of many diversity sensors and the corresponding technical intelligence, to be able to react in an adequate way. The classic recognition methods are mostly based on analyzing camera pictures by means of pure mathematical algorithms. In contrast to this the brain works in a completely different way. A human brain memorizes "images" all the time and associates them, if the human body wants to see something. That means, that the optical system for example, does not "scan" the outside world like a photo camera by computerizing the incoming picture. The human being is only able to "see" characteristic values of forms and surface areas with the optical systems, which in turn associate the memorized "images" to recognize the outside world and to memorize the new manipulated "images", should the occasion arise. So, the human being only sees a mixture of computerized images of the "inner world" and the incoming data which associate with the images of the "inner world". However, the human being does not only memorize optical Images, but also images of all other senses: sense of smell, of touch, of hearing, of taste and of pain. This whole scientific theme, concerning how we transform the pure sensor data to images and scenarios, is the topic of the research projects ARS PC (ARS Perceptive Consciousness). The second research project (ARS PA: ARS Psychoanalysis) deals with the question, if the Id-Superego-Ego model of Sigmund Freud can be taken over for technical concepts. The nature has beyond doubt achieved great success, if one compares the ability of a human being in opposite to animals.. This model is much more complex than all other models, which have been used in automation up to date, like control systems based on methods of neural networks, fuzzy logic or of knowledge-based systems. Two instances, the Id and the Superego, are operating separately and often diametrically to one another; meanwhile the EGO should realize the optimal and most efficient solution in the outside world. In this function network feelings have a decisive roll, and of course it is important which parts are conscious and which remain in the subconscious space. The technical world has not dealt with such questions up to now. To summarize, if we want to increase the performance of control systems dramatically, to handle complex processes, we should take over the achievements of the NPP - especially those of the last 10 or 20 years - and face each other positively/with a positive attitude.

Personen

Projektleiter_in

Projektmitarbeiter_innen

Institut

Contract/collaboration

  • Seibersdorf Labor GmbH

Schlagwörter

DeutschEnglisch
PsychologiePsychology
Komplexe Systemecomplex systems
Symbolisierungsymbolization
Künstliche IntelligenzArtificial Intelligence

Publikationen