A simplified route choice model using the shortest angular path assumption

Route choice within transportation network analysis is generally assumed to correspond to a utility function. Although cognitive factors are known to affect this function, most route choice models assume that minimum time or distance are the primary elements affecting the chosen path. However, evidence from cognitive science suggests that the number and angle of turns during a pedestrian journey affects the perception of its length. Does this distorted estimation affect drivers’ route choice? In this paper, we build a model which assumes drivers will try to take the minimum angular path rather than the more usual minimum block-distance path. In order to concentrate on the cognitive choices involved, we use a simplified model. We form a network from standard road-center line data. The lines comprising the network are treated as a graph, where each straight-line segment acts as a node, and the edge-weight to any other segment is the angular turn to it from the current segment. To allocate trips, every segment is treated as a both a possible origin and a possible destination, and we assume the shortest angular path between the two is followed. This allows us to use the standard graph measure of betweenness to act as an approximation of relative vehicular flow through segments. We validate the model against daily traffic-flow rates collected for a small region of London in a previous study. We find that the minimum angular path model correlates strongly with observed traffic flows, and that it significantly outperforms a minimum blockdistance model.