Site icon ASCE's 2021 Infrastructure Report Card

Saving Our Infrastructure and Cutting Expenses: Artificial Intelligence and Predictive Road Maintenance

Guest Blog by EyeVi Technologies

The United States infrastructure report card is nothing to write home about. According to the 2021 ASCE Report Card for America’s Infrastructure, the overall score for America’s infrastructure is a C-. This more than imperfect grade is reflected well in the satisfaction rate of amongst everyday Americans with their infrastructure. In a recent survey by Ipsos, Americans are slightly more satisfied with the state of their roads and bridges than Spaniards. However, South Africa, Belgium, and Poland are more satisfied, with China topping the list. (figure 1)

Figure 1: Proportion of satisfaction (very/fairly satisfied) scores by country. Adopted from Ipsos 2021

In accordance with the Ipsos survey, the report card also shows that one of the biggest factors that bring our grades down is road infrastructure; specifically – 43% of major US public roadways are in poor or mediocre condition. That estimate doesn’t even encapsulate local roadway pavement conditions, which are often significantly worse than major U.S. roadways, due to lack of funding. Overall, the backlog in road maintenance amount to billions of dollars, and the costs continue to rise.

Meanwhile, current manual data capture methods of pavement conditions waste further time and resources, as running a full review that is both time-consuming and expensive. A lack of sufficient data in turn leads to difficulty in prioritizing the most urgent road repairs and maintenance of asset items.  Fast data capture and automated data processing are urgently needed to capture much-needed condition information and help states and localities prioritize.

High-quality equipment for fast data capture

Fortunately, technological advancements are better allowing for quick and cost-effective data collection. Instead of driving road by road and manually recording defects and the state of asset items, technology is enabling various strategies for capturing geospatial data, ranging from cheap and simple smartphone-based solutions to expensive high-tech devices. There are choices for everyone according to their needs for data quality and availability in terms of budget.

However, to ensure sufficient image quality and location accuracy for automated defect detection, the hardware should include a 360° panoramic camera, a LIDAR scanner, and a GNSS/INS system (figure 2). EyeVi offers such data capture services in addition to AI-powered defect detection software. Using hardware that can be mounted to the roof of a car makes data capture fast and safe. The data can be gathered while driving at a normal speed and there is no need to inspect the road visually – this work will be done later by artificial intelligence.

Figure 2:  A car with mountable high-quality mobile mapping equipment (left) capturing the data (right)

AI-powered digitalization of the captured data

When the data is captured, the next step is to digitalize it, i.e., annotate the data for objects of interest and process it (figure 3). There are various AI-based software and data processing platforms that offer this kind of service. Using artificial intelligence in various domains has been on the rise for the last decade. Road infrastructure management is also a field that can benefit enormously from this high-tech solution.

First, the data processing is automated – no more manual annotations. Second, the quality of the data is consistent throughout the datasets – there is no subject bias in the annotations that tends to result from human-annotated data. Third, using AI instead of a whole team of annotators makes the data processing faster and cheaper.

Figure 3: Point cloud image of the captured data (left), base data (middle), and AI-based annotation (right)

Smart data management saves the grade

Combining the two methods – fast data capture with hardware that includes market-leading sensors, and AI-powered data processing – leads to sustainable road infrastructure management. How does it make management sustainable? Investing in the use of an AI-powered and automated 3D mapping platform enables road consultants and asset managers to have a constant overview of the physical road network in maximum detail. Road defects, attributes, profiles, road signs, safety barriers, markings, crossroads, and so much more – everything necessary for predictive road maintenance, road asset and traffic management, safety auditing, and autonomous transportation. The result? Road maintenance backlogs start reducing and costs go down as the priorities get set and resources are directed where needed the most in a timely and efficient manner.

With EyeVi’s market-leading sensors and AI-powered digitalization coupled with smart, automated data management, there’s no reason why we can’t pass our next test with flying colors.