Psychology of Route Choice in Familiar Networks: Minimizing Turns and Embracing Signals

Abstract

It is widely accepted that path choice of a trip is dependent on trip characteristics, network attributes, and a traveler’s personal characteristics. The best-known network variables that influence route choice are travel distance and travel time. This research attempts to study the influence of other network variables, namely signals, turns, and roadway classification on route choice. Real-world trip data from path trajectories tracked by a global positioning system (GPS) in an urban area are used to isolate nearly 5,700 unique real paths. Procedures to compute the theoretical shortest time path (STP) and shortest distance path (SDP) based on travel time and distance as impedance variables, respectively, are developed. Street network data are augmented with data on signalized intersections. Procedures to identify turns and road classes along the real and theoretical paths, and methods to quantify turn penalties are developed. The real paths are compared to their STP and SDP counterparts to identify discernible relationships between the network variables and the path choice. The number of traffic signals along the path is found not to be a statistically significant factor during the path selection process. In contrast, for trips shorter than 16.1 km (10 mi), drivers embraced more signals along their chosen travel paths than the number of signals that are present along the SDP or STP. It is observed that drivers are willing to spend longer time or travel longer distances on paths that have fewer turning movements. Furthermore, there is statistical evidence to indicate that real paths have fewer turns per 1.61 km (1 mi) than both STP and SDP. When they must make a turn to complete their trips, drivers seem more prone to making the turn at a signal-controlled intersection, while at the same time trying to minimize the number of turns occurring at nonsignalized intersections. Most notably, for trips shorter than 8.05 km (5 mi) in length, real paths have a statistically significant fewer left turns per 1.61 km (1 mi) than right turns per 1.61 km (1 mi). This leads to the conclusion that drivers tend to minimize left turns while selecting a path. Exceptions to these observations are very few and they all happen with paths longer than 16.1 km (10 mi). It is anticipated that the findings from this research will be influential and make it easier to find paths that are more consistent with drivers’ real choices and consequently provide more sound solutions to modeling transportation networks.