Field experiments with an intelligent leaf sensor

This paper describes an innovative sensing method which has been developed to replace leaf sensors in plant propagation systems. Such chambers present critical situations for control because of the high humidity levels to be maintained, which makes direct sensing methods unsuitable. The method of humidity application involves psychrometric processes that create large time delays and renders feedback only control inoperable. The new method implicitly relates the transfer phenomena between the plants' microspace and the environmental macrospace, through the selection of control parameters (net weights). An adaptive system guides the selection of control parameters on-line, conducts real experiments on the site and uses a modified descent method to maximise performance. The performance surface was studied and a modification of the searching algorithm has improved the learning rate significantly.