Is a suggested traffic control currently in operation at any specific location?
Xuanwu Chen: Yes.
Mohammed Hadi: We are about to finish the delivery of the final report in the near future. What we did is compared the developed method with the real-world observations, and we found out that the method is producing reasonable results. So the plan is to implement this on additional intersections with the intent of some possible fine-tuning of the signal timings.
The Compliant to Queue warning messages: are they relayed to connected vehicles only? If so, how does your model predict the driver behavior and the actions to respond to such messages?
Samaneh Khazraeian: The answer is no: it’s not only connected later. The model—the Queue warning system—by changing a certain percentage, can be connected or not connected. So the messages are going to disseminate through the dynamic message signs or onboard units.
Mohammed Hadi: In that case, it’s based on the assumption that the agency wants to continue to have the infrastructure dynamic message signs. Then you can assume all the drivers will have access to those dynamic message sign messages. If the plan is to only depend on connected vehicles, then one can test the messages with just connected vehicles receiving it. The compliance is a Monte Carlo simulation where you can select the people that complied based on the Monte Carlo simulation. That would require different random numbers, and you take the average based on that, and you see the variation.
What do you use for the baseline car feature expectations for connected vehicles for your market penetration analysis?
Md. Shahadat Iqbal: All the new cars will have the feature. From the very first tier, you will have 100%. All the new cars will have the features.
Mohammed Hadi: This assumption can be varied. The idea is to have the effect of socioeconomic factors on the market penetration between regions and between zones within the region, but the assumption itself can vary. Our assumption was the first tier after the NHTSA mandate, you would have 100% of all new vehicles to have connected vehicle equipment, but that could all change. There is some discussions about only 50% by 2020 or 2021. When the mandate will become active, you may have only 50% of new connected vehicles with connected vehicle equipment. Our method is independent of that. You can set it the way the expectation is with an installation of the equipment. The main contribution is that we are connecting the income and the car ownership in the region to a model between regions, between zones within regions, and between links within regions, because assuming a fixed value of market representation on all links in the region may not be correct, and that’s the main idea.
Can research be used to develop traffic signal performance measures?
Mohammed Hadi: Yes. We produced a dashboard that can be used by agencies to look at it so some of the indexes that Xuanwu (Chen) talked about are, in fact, produced and combined in the dashboard that the agency uses.
How and when are these principles going to be tested in a real-world environment, and where did you get your funding?
Mohammed Hadi: As you know, a lot of funding comes from the U.S. DOT, but states, in fact, have already started investing. I know that in Florida we have started investing in connecting vehicles, both at the state level in Tallahassee, the district level, and even the county here in Miami and Broward. They are starting to be very interested in these technologies. I’m thinking that you will see that the funding will be there to start testing these, and for when these principles will be tested, we know U.S. DOT will have three pilots—in Tampa, New York, and Wyoming—and they’re testing several applications of connected vehicles, and FHWA has announced funding for additional projects to do emerging technologies to support active traffic maintenance, and they’re going to announce new ones in the near future. We have the Smart City effort as well. A lot of these concepts will be tested in the near future—in the next two or three years—at the time this mandate will be there, and we will have connected vehicle technologies. We know, in 2017, some of the Cadillac CTX will have connected vehicles in them, so by the time these vehicles are there, a lot of these concepts will be ready to be tested in the real world.
For Xuanwu, how was the threshold generated?
Xuanwu Chen: We started the threshold using 2.5 for the travel time index and 0.5 for the maximum ratio at that time, and it generates the results with the field observations. We also tested the sensitivity of the thresholds so we can understand how the threshold affects the framework. The threshold is based on the expectation of the field conditions you want to achieve. So sometimes, if you have a high threshold, you have little accommodation to change the signal timing. But if you have a lower threshold, it’s hard to achieve that recommendation.
Mohammed Hadi: It’s really based on trial and error. You look at the corridor, and you see that setting the threshold at a certain value produces good results. In fact, it was very clear to us that a certain threshold produced good results as we went to do field confirmation to confirm the results. So in that corridor, it made sense. Now, in the next corridor, maybe we need to come in an automated way and have sensitivity analysis on it—we will try to do some optimization of those thresholds to figure that out—but it’s also related on how much emphasis the agency wants to put on thresholds, so it is a combination of judgment and sensitivity analysis. There may be opportunities for optimization in the future.
United States Department of Transportation
