Turchenko V. Neural network-based methods and means for improving the effectiveness of distributed sensor data acquisition and processing networks

3. Research object - distributed networks for sensor data acquisition and processing. The purpose - development of neural-based methods and creation of software-hardware means of their implementation, devoted to improvement of effectiveness of sensor data acquisition and processing in distributed networks. Researches methods - structural and functional analysis, neural network theory, simulation modeling. Theoretical result - neural-based methods of prediction of sensor data drift in order to increase time between failures and structure of the network, which provides required processing speed. Practical result - controller for neural network calculation of the middle level, software of the higher level. Economic effect - decreasing the cost for calibration/testing, using low cost hardware and software. Areas of usage - industry, scientific recearches.