Modellierung navigatorischer Fahrerpräferenzen für die personalisierte Routenberechnung

Auteurs

• Andreas Landau

Paramètres

• Manuels
• Manuels universitaires

En savoir plus sur le livre

The routing component of modern navigation devices relies on shortest path calculations using an objective function and edge attributes from an underlying graph. Commonly, these calculations aim to minimize route duration or length, aligning with general driver preferences. This work enhances routing by personalizing it according to different driver types. The objective function and edge attributes are automatically adjusted to reflect individual preferences. To accurately model these preferences, three models of driver behavior are developed, addressing both time and space dimensions. The first model focuses on observed speeds on edges, while the second provides a measure of a driver's mobility network coverage across all edges. The third model evaluates routes by analyzing paths of edges through both dimensions. Correlation analyses confirm that results from these models align. A robust two-dimensional model is then established using well-known attributes to classify drivers based on their navigation preferences. This model facilitates the personalization of both the objective function and edge attributes, which are evaluated against standard routes through various measures. The entire modeling and personalization process is implemented on a large-scale data set, quantifying improvements in navigation devices for individual drivers and the broader sample classes.

Achat du livre

• Année de publication: 2016