Acquiring knowledge and numerical data to support CBR retrieval

This paper illustrates a Knowledge Acquisition and Representation tool (KARM) to support the design in a restricted and specific domain (i.e. the design of tyre treads for motor racing). The main goal of the KARM tool is to allow the acquisition and representation of track knowledge in terms of its morphology, and meteorological features. Two ways to analyze the geometric structure of tracks are combined into KARM; the first one is based on the acquisition and representation of qualitative knowledge (the block view technique), while the second one is based on the analysis of quantitative knowledge (P-Race Telemetry). Moreover, two fuzzy based modules that allow the representation of uncertain and imprecise knowledge about weather and asphalt conditions will be pointed out.