CHAPTER 12

Automated Pavement Maintenance

12.1 Description of Automated Pavement Maintenance Applications

Two different types of automated pavement maintenance applications were found in the literature: automated pavement data collection and automated pothole repair.

12.1.1 Automated Pavement Condition Collection

Traditional pavement condition monitoring has involved visual inspection and in situ tests to assess road distress, unevenness, rutting, and crack width and depth. UAVs have been piloted to capture and process images, create digital 3D road surface models, and extract features and measurements (Outay et al. 2020). Li et al. (2018) used high-resolution videos and point cloud generation to detect potholes. Most applications are still in the development phase (Zhang and Elaksher 2012, Brooks et al. 2016, Shaghlil and Khalafallah 2018).

Two different vendors have created products that utilize connected vehicle data to assess pavement condition. One vendor uses data collected from connected vehicle sensors and cameras that are analyzed in real time and uploaded to cloud servers anonymously. Through artificial intelligence technology, roadside markings, signs, and traffic lights can be visualized and identified for potential damage. The system provides the location and severity of roadway damage and can provide information on roadside assets (Stone 2023). Another system can detect potholes using the dash cams from a network of more than 400,000 vehicles. Machine visioning can be used to interpret these images to identify the presence and severity of cracking, striping conditions, and the presence of potholes to generate a pavement condition score. The system can automatically detect cracking and potholes within approximately 1% of the accuracy of lidar. Detection can be done in real time (Blyncysy 2023).

12.1.2 Pothole Maintenance

Two companies have developed automated machines to fix potholes. The systems developed by both companies operate in a similar manner. The machine cuts an area around the pothole into a square, which makes the repair easier. A brush then cleans the area and collects the cut asphalt for recycling. The machine then fills the hole. For one of the machines, an average pothole can be filled in 8 minutes, which is much faster and less labor-intensive than manual repair. This machine is currently being used in the UK (Holderith 2021). The other machine uses a similar process and has the ability to fix a pothole within 5 minutes. The difference between the two systems is that the second machine has a larger capacity to carry asphalt plugs and fill material.

12.2 Examples of Automated Pavement Maintenance Applications

Information about how agencies have utilized automated pavement maintenance applications was gathered from a review of the literature and a survey of agencies.

12.2.1 Responses from the Survey

A survey was conducted to gather information about the automated processes that IOOs have implemented or are planning to implement, as described in Chapter 3. Agencies were asked about the automated processes that they have used or piloted and the processes that they thought could benefit from automation.

Around 68% (n = 21) of responding agencies are using some types of automated processes for pavement data collection, with 13% (n = 4) planning to use or evaluate the use of automation for this purpose. Around 19% (n = 6) of agencies are not using or planning to use automation for pavement data collection. Additionally, 9% (n = 3) of agencies noted the use of some automated processes for pavement repairs such as fixing potholes, with 25% (n = 8) indicating that they had plans to use automation for pavement repairs in the future. The majority (66%, n = 21) of agencies were not using or had no plans to use automation for pavement repair.

12.2.2 California

Caltrans, along with a vendor, introduced a semi-automated machine after the death of two workers who had been patching potholes. The system could be operated without exposing workers to traffic and was used in the Caltrans District 4 region to pour 5 tons of asphalt within a very short period. However, the use of this system was not continued due to some safety and handling issues (Bennett and Velinsky 2014).

12.2.3 Georgia

GDOT conducted a study to derive a workflow for pavement maintenance involving the use of uncrewed aerial vehicles. The proposed method can potentially detect anomalies in a pavement surface that can indicate the presence of possible subsurface voids in the pavement structure. The study also found that red-green-blue and infrared images can be used to detect pavement defects more precisely than other methods (Irizarry and Rakha 2022).

12.2.4 Kansas

The Kansas DOT uses a system to continuously collect pavement deflection and other surface-related data. These data can be used to determine the structural capacity and remaining life of the pavement. The system can also identify potential sinkholes that are not detectable by other methods (ARRB Systems 2023a). This system is also used by ITD (ARRB Systems 2023b).

12.2.5 Other Automated Pavement Maintenance Applications

Researchers from the University of Liverpool in the United Kingdom invented a technology to identify and treat potholes automatically. The patented technology uses high-quality laser-scanned data and an algorithm to detect millimeter-sized cracks. The system can then fill these cracks with suitable materials without decreasing the skid resistance of the road. The overall system can also prioritize work locations and schedules by predicting the future condition of the

identified defects. This project was funded by the United Kingdom’s Department for Transport (Df T), and the resulting technology was commercialized by a vendor (Thomson 2020).

The city of Ottawa, Canada, piloted the use of an automated pothole repair system to improve the efficiency of winter-affected pavement maintenance. The machine blows air into a pothole to clean the area and then pours hot-mix asphalt to fill the pothole (CBC News 2019).

12.3 Summary of Applications for Pavement Maintenance

Three states have piloted several automated pavement maintenance applications. One state tested the use of a semi-automated machine for pothole repair, one used UAVs to detect anomalies in pavement surfaces, and one used an automated system to continuously collect pavement deflection and other surface-related data.

12.3.1 Advantages

The use of autonomous and semi-autonomous systems for pavement maintenance offers a number of benefits:

• Reduced costs related to time, maintenance, and labor.
• Improved safety for workers by eliminating the need for traffic intervention.
• Reduced error in characterizing pavement surface conditions.
• Optimized maintenance through real-time life-cycle analysis of pavement condition data.

12.3.2 Disadvantages

None of the technologies were sufficiently advanced to fully assess disadvantages. However, California’s experience with safety and handling issues suggests a major disadvantage for the technology.

12.3.3 Costs

Most of the sources consulted did not provide specific estimates of cost.

12.3.4 Status

All of the applications for automated pavement maintenance are in their infancy. As a result, none are widely available.