Ozone is one of the main air pollutants with harmful influence to human health. Therefore, forecasting the ozone concentration and informing the population, when the air-quality standards are not being met, is an important task. In this international conference paper, an on-line updating, or evolving, GP model of ozone concentration is evaluated. Its main advantage is the ability of learning from scratch, i.e. with almost no prior knowledge or data. This means that it can be used for modelling various variables shortly after the measurement station is established. To assess the viability of the evolving GP model for the ozone forecasting it is compared to the full GP model and multilayer-perceptron ANN model. The investigation shows that the evolving GP model performs sufficiently good to use it for informing citizens about the possibility of high and alarm concentrations occurring, especially, if the prediction-confidence interval is taken into account.
B.03 Paper at an international scientific conference
COBISS.SI-ID: 28052519Neural network model was implemented on an embedded computer, which had less computational capabilities than regular personal computers. Embedded computers are important in applications where constraints of electric power exist or they operate in extreme environment, e.g., extremely low and high temperatures, high relative humidity. This implementation enables online use of the model in the environment where measurements take place. This implementation is now being tested with the model of diffuse solar radiation at the regular meteorological station. The next step is the use of this embedded-computer implementation with power consumption less than 2 W for online forecasting of ozone pollution at places where measurements take place.
F.06 Development of a new product
The paper in the popular daily newspaper Delo describes the significance of ozone-pollution forecasting for planning activities in countryside and for reduction of health damage caused by ozone. Ozone itself, its negative influence on human’s health, the procedure of ozone forecasting, the estimation of air quality and publication of data on a public web site http://www.kvalitetazraka.si are described for a general public. The purpose of the publication is informing the general public about the new media that makes forecasts of air pollution in Slovenia with ozone and other pollutants available online more comprehensively and in greater detail. Major enhancements are two new indices available for entire Slovenia at spatial resolution 12 km x 12 km and time resolution of 1 hour: AQI index that shows the impact of ozone on air quality and health and AQHI index that shows the impact of air quality on health.
F.18 Transfer of new know-how to direct users (seminars, fora, conferences)
COBISS.SI-ID: 27973415