The necessity to improve the quality of raw materials and the necessity to use computerized and computerized systems, able to monitor production process real time parameters, is pressing for expanding companies. The high unevenness of raw materials, especially when they are subject to be processed under production-chain phases, is demanding more refined control systems. Control systems must be fast, linear and easy-fitting to situations that could slip through mathematics process modelling actions.ITACA project is proposing to research and develop ad hoc supporting control system to improve food production chain of Piedmont Lavazza and Ferrero,flagship of Italian food industry. They have to manage a huge quantity of data, regarding the conversion process, many different nature parameters take part, such as color, dimension, chemical markers.
Companies need to use these information in lock step, so they should be able to reproduce under models the same process running, just as human brain does. We need information that cross the limitations of classic statistics and so ITACA demands artificial neural networks use, data driven systems employment.
Neural systems shall be used to increase production process stability. This will be possible for better parameters and conversion project monitoring and this will provide steady, very high quality of end product. One of NLG goals within ITACA is to be able to develop self diagnostic system. This system should adjust process parameters real time, as requirements demand. Several neural models shall be projected. All of these shall be adjusts to conversion processes of coffee and hazelnuts.
A neural nets based platform shall be projected and it will involve fuzzy system and statistical models as Hidden Markov Models (HMM) too. This models can elaborate parameters time series and assess real time situation and to forecast close time evolution of production process main parameters.How getting along? Numerical time-series process parameters will be generated. Ad hoc treatments shall be performed according to time series parameters elaboration results. Image processing techniques, chemical analysis and NIR will be taken into account.
ITACA project constitues a multidisciplinary field of work. NLG on behalf of Politecnico will join in ITACA with other working realities, as Lavazza and Ferrero, as CMC, manufactures and servicing of industry grindstones and rollers and as GBV, building industry machines to transform hazelnuts. We will study parameters linked to grinding cogs, linked to roast coffee with CMC and with GBV, we will study taste sense perceiving with Lavazza and Ferrero.