BIOMASS ESTIMATION IN PLANT-CELL CULTURES - A NEURAL-NETWORK APPROACH

The special characteristics of plant cell cultures make it difficult t o use conventional analytical techniques for on-line biomass monitorin g. Meanwhile, promising results have been obtained using mathematical models and recursive estimation algorithms. However, in this case, add itional experimental effort is necessary to obtain a reasonable descri ption of the process. Recently, techniques using more empirical approa ches have been proposed to describe complex processes, minimizing the experimental work needed for their application. In this paper, we repo rt on the use of artificial neural networks to monitor biomass evoluti on in plant cell cultures. The results obtained with a three-layered n etwork are presented. Method requirements and capabilities are compare d with the method based on the extended Kalman filter used in previous work.