This paper reports on models developed from data collected using the PARKIT parking choice simulator. PARKIT provided an experimental environment in which drivers' choice of car parks, and of the routes chosen to reach them, could be observed and the influence of different levels of parking-stock knowledge (derived from experience or from information provided via roadside message signs) monitored. Separate models were estimated for the drivers' initial choice of car park and for their revision of that choice as their journey progresses and they learn about actual conditions. The importance of price, walking time and driving distance is confirmed but the addition of variables describing the drivers' choices on previous days, their expectations and their immediately preceding route-choice, greatly improved the models' explanatory power. It is noted that variables such as these are not generally considered because they are rarely available to the modeller. Different discrete choice model structures were found to be appropriate for different decisions. Route choice was represented as an exit-choice model (whereby each journey is treated as a sequence of decisions -one at each intersection encountered). The paper discusses the incorporation of these choice models into a network assignment model and concludes that much of the power of the choice models is lost if the network model is not able to support use of information about travellers' socio-economic characteristics and knowledge of the network and about the detailed network topology.

No abstract

No abstract

Content-based indexing and retrieval of images and video requires a proper semantic description for image content.This paper discusses the mapping of high-level, application-specific, features to the visual primitives that are accessible through image processing techniques. A major difficulty is that there are currently no established methodologies for describing the contents of an image in terms of semantic features, and we suggest that semiotic approaches could be adapted to the task of image description in limited areas of expertise. After reviewing current trends in the mapping of high-level to low-level features, we present preliminary results that suggest a possible strategy for mapping semiotic descriptions to image processing primitives.

No abstract