Roadway Sensors Provide Real-Time and Historic Data
Robert Bray is a senior software architect at Autodesk, Inc., a computer aided design firm that readers may have heard about. He recently sat down with Informed Infrastructure to discuss his view of sensored roadways, and the uses of the information they develop.
Informed Infrastructure: Robert, can you tell us how you came to be interested in sensored roadways?
Bray: At Autodesk, I’m primarily responsible for the Infrastructure Suite, including our roadway and utility design products, and that’s where my interest in sensor data comes from. It’s true that the current focus of sensor data is on the real-time use of the data to manage traffic flows. But we’re learning that the historical data being gathered is also incredibly valuable. That huge volume of data about traffic speeds and volumes, and the way they change over time, can be leveraged to do much more effective planning based on real data, not guesswork. So that’s my interest; how can we manage this data, what can we learn from it, and how can we build that knowledge into our design tools?
Informed Infrastructure: What is the current state of sensored roadways, and how have things changed over the last ten years?
Bray: Well, ten years ago there really wasn’t any sensor data to speak of, but starting about five years ago sensors started to appear that fed data to commercial mapping systems and that could be used to forewarn drivers about traffic conditions. Today, the installation of sensors in urban areas is accelerating. Sensor technology on the hardware side is way out in front of the software required to process all the data they collect. Today they tend to gather data on traffic volume and speed, but the quantity of sensors—and sensor data of course—is going up rapidly. And I think we’ll see more types of information being gathered eventually. When sensors are built into new roadways, for example, they may be able to gather information on weight loads. In some ways, the use of sensors is still in its infancy.
Informed Infrastructure: What do you realistically foresee being implemented over the next ten years?
Bray: I think you’ll start to see a much larger investment in sensors for urban areas, and more use of historical data to make driving more efficient. Planners will have a much better idea of what’s needed. If you can reliably predict peaks in volume, for example, planners can find ways to increase lanes in one direction at certain times of day, or widen roadways, or even build new roads to offload traffic. This is what planners have always done, of course, but large volumes of data gathered over years is much more useful than data gathered by a person sitting in a car at the side of the road.
Informed Infrastructure: Would you say there is a clear picture of what changes are desirable, or is there some controversy about what’s needed?
Bray: I don’t know that there’s any controversy, and it seems clear that sensors add value. But I do think it’s up to technology companies to drive innovation to show what’s possible and come up with new uses for the massive amount of information that’s being developed. In turn, that will drive the addition of more sensors, and probably more sophisticated sensors. As I’ve said, my primary interest is planning, but I think it’s completely realistic to manage traffic in real-time with sensor data to prevent traffic jams and keep traffic moving more efficiently. There are some interesting test beds for that now and, longer term, that could easily be the future of driving. Traffic congestion in urban areas will continue to be a problem, and that applies a lot of pressure to adopt new solutions.
Informed Infrastructure: Are there technical challenges yet to be resolved?
Bray: Speaking as a computer scientist, my real interest is in the sheer volume of data. If you think about the number of cars that pass along a given stretch of roadway in a densely-populated urban area, and the fact that data is gathered continuously, 24 hours a day, every day… well, you can see that the quantity of information quickly becomes so huge that analyzing it is a challenge. So I think cloud-based analytics will become a dominant solution in this area. We’ll also start to see heuristics and patterns start to emerge from the data, and the science of traffic analysis will become more sophisticated. Eventually we may be able to build very sophisticated predictive models based on factors like the numbers of businesses and residents in a downtown core, and there will be much more value gained from investments in better planned and designed roadways.