his project attempts a systematic study of the application of artificial intelligence tools for the modelling and optimisation of systems based on electrochemical sensors. When these have a simple response, such as in the potentiometric case, a simple modelling employing artificial neural networks (ANNs) is sufficient. When the response is much more complex, such as in the voltammetric, dynamic response or impediometric cases, a pre-processing is called for. This is used for extracting the most significant information previously to the modelling. The Wavelet transform is the intended tool for this pre-processing. To ease the generation of information and to automate the operation of the proposed systems, these will employ flow manifolds based on the SIA technique. Hence, a lrge number of reference solutions can be easly generated as mixtures of standars; also the handling of samples is facilitated in the same way. The configuration of the RNA with best modelling ability, traditionally found by trial and error, can be optimised with tolls like genetic algorithms. The proposed analytical systems, known as electronic tongues, are intended, not only for multianalite determinations, but also to cases where the information sought is not directly related to chemical concentrations. This happens when the goal is to classify varieties, to qualify a perception or a situation, per example, toxicity or bad taste. The most applied part of the project employs the specified concepts for the monitoring of chemical parameters during advanced water treatment processes, such as swimming pools equipped with ozone treatment. As an added contribution, the global management of these facilities will be attempted employing a computer system, incorporating climate and chemical information, and including a social diffusion plan of these new technologies
|Effective start/end date||13/12/04 → 13/12/07|
- Universitat Autònoma de Barcelona (UAB)
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.