Journal of Computing and Information Technology - CIT 16, 2008, 4, 279–286 doi:10.2498 /cit.1001395 279 An Integrated Knowledge Base for Modelling and Predicting Vehicle Real-world Emissions as a Function of Driving Behaviour Kinematics Mario Rapone † , Livia Della Ragione and Giovanni Meccariello Istituto Motori CNR, Naples, Italy A multivariate modelling approach was developed by Istituto Motori to model and predict vehicle real-world emissions. Complex driving kinematics is represented by two blocks of variables, which require the development of a hierarchical multiblock emission model, where the two blocks of variables represent overall and instanta- neous features of each driving cycle associated to a trip. The multiblock model was applied to analyze and model emissions of the large database built in the ARTEMIS project. In this database we collected emission measure- ments performed in European laboratories relative to real driving cycles which are statistically representative of many European traffic/road conditions from congested to rush-hour traffic in urban, rural and highway roads. Data concern a varied fleet of vehicles differing in technology and class. To develop a tool useful for mobility analysts for traffic environmentalimpact assessment, a knowledge base was envisaged to integrate the data warehouse and the model base to build a user interface for driving cycle kinematics and emission analysis. In this paper the modelling approach is presented together with overall emission and driving kinematics characterization based on experimental results, as well as functional analysis of the knowledge base structure and the information tool. Keywords: hierarchical PLS multiblock, kinematic pa- rameters, decision support systems (DSS), real-world emissions, data warehouse 1. Introduction The effect of greenhouse gases on climate has gained global interest and pushed top level pol- icy and decision-makers to promote action for the reduction of emissions. Kyoto 2008 will define scenarios, policy and strategies for the control of CO 2 produced by each country. Transport activities contribute significantly to air pollutant emissions. Besides CO 2 ,other traf- fic emissions have a major impact on human health, and for historical towns, on architectural heritage. Evalution of emissions has therefore gained institutional importance in the interna- tional community. Emission evaluation can be performed for dif- ferent purposes and uses at different space and time scales within trasport envinronmental im- pact assessment. National, regional, and city emission inventories require overall average emission estimation based on a few input pa- rameters such as fleet composition and average mission profile, average speed related to av- erage traffic situation (Macro scale analysis). Development of regional and city transit plans require detailed representation of the area traf- fic and transit network into links. For each link average emissions are generally evaluated for each individual vehicle as a function of vehi- cle flow and average speed (Meso scale anal- ysis). More detailed analysis requires vehicle driving behaviour represented by the time series of vehicle speed on the link, hereinafter called driving cycle (Micro scale). The assessment of the environmental impact of measures regard- ing road crossing control or development of new infrastructure requires analysis of instantaneous emissions of individual vehicles to estimate the effect of the planned measures on vehicle speed and acceleration.