Recurrent Neural Networks for Moisture Content Prediction in Seed Corn Dryer Buildings

Conditioning seed corn is a short, yet crucial, portion of the seed production process. Seed corn must be conditioned prior to removing the seed from the cob to prevent damage, requiring constant monitoring by farmers. This paper evaluates the use of an echo state network for the prediction of seed moisture content and compares it against an Elman network. The results are determined to be good enough for inclusion into a commercially available dryer monitoring system.