Adaptive route selection for dynamic route guidance system based onfuzzy-neural approaches

摘要

One functionality of an in-vehicle navigation system is routenplanning. Given a set of origin-destination (O/D) pairs, there could benmany possible routes for a driver. A useful routing system should haventhe capability to support the driver effectively in deciding on annoptimum route to his preference. The objective of this work is to modelnthe driver behavior in the area of route selection. In particular, thenresearch focuses on an optimum route search function in a typical in-carnnavigation system or dynamic route guidance (DRG) system. In this work,nwe want to emphasize the need to orientate the route selection method onnthe driver's preference. Each feasible route has a set of attributes. Anfuzzy-neural (FN) approach is used to represent the correlation of thenattributes with the driver's route selection. A recommendation or routenranking can be provided to the driver. Based on a training of the FN netnon the driver's choice, the route selection function can be madenadaptive to the decision making of the driver