摘要: |
Rapid population growth in Arizona has created several large residential areas that rely on the State highways to provide their primary, daily commuting route. When these commuter routes cross an at-grade railroad crossing, a train passing during peak traffic hours often causes severe congestion. State resources are inadequate to provide flyovers for all of these train crossings and their numbers are forecast to increase. The safety and congestion problems arising from these commuter at-grade crossings are the focus of this research. A study site was selected, train and traffic data were collected, a microscopic traffic simulation model was prepared, and an Early Warning System (EWS) algorithm was developed. The EWS algorithm gives 'extra' green time to (train) conflicting traffic movements before the train arrives, taking the time from the other movements. Five cases were studied, each having two to six scenarios. Four major variables were studied: (1) crossing gates down time, (2) length of time the MOEs were collected, (3) conflicting movements traffic volumes, and (4) predicted arrival time error. The EWS algorithm was also successfully programmed into a NEMA controller using Hardware-in-the-Loop to couple it to the simulation model. Four generalizations are tentatively supported by the results but additional site studies are required for verification. First, the complex dynamic interplay of geometrics and train and traffic volumes makes the EWS effectiveness highly site dependent. Second, there must be enough pre-train vehicles present on conflicting movements that derive delay improvement to overcome the increase in delay to the other movements. Third, for safety reasons, an increase in overall intersection delay caused by the EWS may be justified to reduce long queues from backing-up into other intersections or onto freeways. Fourth, rather than control signal timing, the EWS may be used to reduce congestion by alerting drivers with a DMS of a train's imminent arrival so they can take alternate routes. While the EWS was ineffective for the study site, the results may have been confounded by insufficient pre-train queue sizes and lack of a single dominant commuter movement (the study site had strong cross flows). A follow-up study is recommended at a site with more favorable geometry and traffic volumes. |