Josse De Baerdemaeker
Department of Biosystems, division MeBioS, Katholieke Universiteit Leuven

Agricultural production is also part of a long chain of activities that starts from the seeding (or even earlier) and stretches all the way to the consumer. They should meet consumer expectations in terms of quality, safety and also value or price. Agricultural and biological production systems also face some challenges posed by the systems characteristics. Agricultural and biological production systems in general can be very variable (natural variability of biological processes) and they are also subjected to many disturbances, most of which are unpredictable. A large effort has already been made to observe the crop or animal performances.
The spatial and time dependent variability of plant production can occur at different temporal or spatial scales. Tools are available to describe the variability, to measure or estimate the status of soils and plants and the effect of a treatment in terms of growth, yield and quality.
The available measurement and sensing technologies yield a vast amount of data that allows the analysis of the variability that exists in crop production. Novel approaches to statistical analysis of these data of time variable processes can support management decisions and help to detect the underlying sources of variability. In this approach the data are analyzed to see if the ongoing process is within some bounds following an expected path. The operator or manager can then look for the causes of deviations and then decide to interfere in an appropriate way. The grower can make decisions on necessary interventions. The speaking plant concept for precision agriculture also implies that control tools and techniques can be applied for optimization of cultivation or environmental conditions such that optimal yields and quality can be achieved even under variable conditions and disturbances. It remains a challenge to decide at what and how many time scales or spatial scales (or combined spatio-temporal scales) the concept of speaking plant for decisions on interventions or treatments should be applied in order to offset the installation and operating costs against potential gains. Nevertheless, the speaking plant approach offers a number of unexploited possibilities for field crop production, especially when combined with synergistic control concepts. Moreover, modeling variability in the production as well as in the post-harvest process can help to make decisions about optimal treatments or about the optimal flow of the products in the markets. Some examples about the use of statistical process control tools for decision making in agricultural production will be given.
Implementation of the tools of monitoring and control of the production processes in the framework of precision agriculture implies that the equipment and the technology used in the field is reliable, both when looking at the mechanical construction and behavior as well as in the operator controls that are available. Only in this way will it be a contribution to the efficient use of natural resources.