Università degli studi di Modena e Reggio Emilia

An instrument for the automatic quantification of glycerol in grapes has been developed. We verify here that this analyte can be taken as a benchmark of a serious disease affecting the grapevines, namely Botrytis cinerea. The core of the instrument is an amperometric biosensor consisting of a disposable screen printed electrode, generating the analytical signal thanks to a bi-enzymatic process involving glycerol dehydrogenase and diaphorase. The full automation of the analysis is realised by three micropumps and a microprocessor under control of a personal computer. The pumps allow the correct and constant dilution of the grape juice with a buffer solution also containing [Fe(CN)6]4- redox mediator and the injection of NAD+ cofactor when the baseline signal reaches a steady state; the instrument leads to automated reading of the analytical signal and the consequent data treatment. Although the analytical method is based on an amperometric technique that, owing to heavy matrix effects, usually requires an internal calibration, the analyses indicate that a unique external calibration is suitable for giving accurate responses for any grapes, both white and black ones.

Our everyday environments are going to be disseminated of devices and sensors that exhibit some degree of autonomy and that collaborate to support users in their activities. This envisions a not-so-far future called "Ambient Intelligence", i.e., environments that show a certain degree of intelligence. Useful abstractions for Ambient Intelligence are represented by software agents, since they are autonomous and capable of acting on behalf of their owners. In the context of the LAICA project we have exploited the agent paradigm to model devices and sensors, and have developed an ad-hoc middleware to orchestrate all the involved components. This paper discusses the adoption of agents in Ambient Intelligence, and explains the design and implementation of the middleware.