Search academic texts

Information × Registration Number 0200U004465, 0199U001057 , R & D reports Title The theory and methods of construction neuronet recognizing systems popup.stage_title Head Pastushenko S.G., Registration Date 14-07-2000 Organization The closed joint-stock company "Basko" popup.description2 The analysis of possibilities to apply traditional approaches and methods testifies to their low efficiency; besides, it is impossible to use them in complex dynamic circumstances when the information obtained is incomplete, unreliable or controversial for solving the tasks of discerning in real time conditions. The project proposes an entirely new approach based on neuronet (neuromophus) concept of discerning systems organization. The project develops methods of formal synthesis for their optimal and quasi-optimal structuring, methods of systems structural and parameter self-organization in neuronet construction. The theoretical results obtained were checked by statistic imitation modelling means. As problem-oriented neuronets, systems of radiolocation discerning and technical condition discerning were chosen. The technical implementation of problem-oriented neuronets was made as imitation programs on modern PCs. Outer environment in modelling is provided by radiolocation and diagnostic portraits. The images of the portraits and their parameters serve as a basis for neuronet structure synthesis; they define the parameters of the classifying element and links between neuronet elements in neuron assemblages. Product Description popup.authors popup.nrat_date 2020-04-03 Close