Measuring the Effectiveness of Advanced Traveler Information Systems (ATIS)

Abstract

The objective of this study was to develop valid methodologies for addressing several limitations of the current Advanced Traveler Information Systems (ATIS) evaluation tools. This study was focused mainly on three enhancements. First, the queue propagation algorithm of the selected tool (DYNASMART-P) was modified to more realistically model traffic congestion. The author proposed the addition of transfer flow capacity and backward gated flow constraints for more accurately calculating transfer flow rate. Second, the study modeled the natural diversion behaviors of drivers who do not receive traveler information. Lastly, statistical models of user responses to traveler information were developed using binary and multinomial logit methods to understand and model the relationship between drivers’ socio-economic characteristics and their responses to traveler information. Among these three enhancements, the first two (improved queue propagation and natural diversion behavior algorithms) were implemented in the enhanced model. The user behavior models, however, were not implemented because their predictive power was not acceptable due to limitations in the data set. The enhanced model was applied to two case studies: 1) verifying the capabilities of the model under a recurring bottleneck scenario on I-40 corridor in the Triangle region of North Carolina, and 2) demonstrating the capability of the enhanced model to measure the effectiveness of U-Transportation (similar to the Vehicle Infrastructure Integration [VII] program in the USA) which has been under development in Korea. The first case study results showed that the improved queue propagation algorithm simulated the bottleneck queue much closer to the real data than the original model. The simulation results also indicated that the actual diversion rate under recurring congestion in the study network was very low. The results of the second case study demonstrated that the enhanced model can evaluate the network impact of new advanced technology in flooding situations and can evaluate the effect of market penetration of the communication technology.