Paving the Way: Using Technology to Preserve Infrastructure

A Little Rock company born of a chance viewing of a television news story is trying to change the way state and local governments maintain roads.

Roadway Management Technologies of Little Rock uses sensor technology attached to vehicles to assess the condition of roads. Developed by Candler McCollum and his former college roommate, William Bates, the company provides live analytics to road maintenance officials in Van Buren and Crawford counties, the city of Cabot and clients in three other states, helping them make key decisions about what to repair and when.

“The key to pavement preservation is applying the right treatment to the right road at the right time,” McCollum said. “There is no way to do that without getting real live data on the roads on a daily basis. What is truly impossible is to watch every single road in a network on a daily basis. The only way to do that is with technology and with crowdsource technology.”

The idea for RMT’s core technology came to McCollum after he saw a news story about the city of Conway hiring a third-party vendor in 2015 to survey and assess road conditions. McCollum, a Little Rock Catholic High School graduate then working in crowdsourcing for Goldman Sachs, thought the third-party survey system sounded inefficient.

McCollum thought a more efficient system should use a crowdsourcing approach. Even a small city has miles and miles of roads, and McCollum didn’t believe that a third party could survey all of them every few years and get an accurate assessment of conditions.

To solve the problem, McCollum, 29, called Bates, his old Auburn University roommate. The collaboration led to the company’s Roadrunner technology, which is helping cities and counties monitor their roads passively and daily through the use of the municipality’s own vehicles at what McCollum said was a fraction of the previous cost.

“I thought about the ability to solve this problem in a different way using big data and machine learning and a crowdsource platform,” McCollum said. “We realized really quickly that no one else was approaching this national crisis, a waning infrastructure, with that type of solution. We started the R&D on building a hardware system that would be able to provide live, real-time data but also educate itself on the rate of deterioration and monitoring different investments into these pavement assets.”

An Arkansas Company

It took McCollum and Bates a couple of years to develop the technology and the company was created in January 2019. In March 2019, RMT received a $50,000 technology development program grant from the Arkansas Economic Development Commission, which allowed the company to start signing up clients.

The company received another $50,000 AEDC grant in 2020.

“We utilized those funds to kickstart the company and really grow it in terms of personnel, etc., starting signing up first customers in summer 2019,” said McCollum, who now employs eight people. “We really are kind of a product of the state of Arkansas.”

RMT signed its first clients in 2019 after testing its technology with engineering companies that specialized in road repair and preservation. For now, RMT operates on the city and county level. Two years after startup, it has clients in four states — Arkansas, Louisiana, Oklahoma and Texas — and McCollum has his eyes on continued expansion. Florida is next on the to-do list, although McCollum is mum about specifics.

“We are doing very well,” McCollum said. “We are growing at a tremendous rate. For a long time, we were really quiet because we were focusing on our ability to scale and execute.”

“After we acquired some customers we slowed down and really focused on the execution with those customers instead of trying to grow too fast and get too big for our britches,” he said. “We wanted to make sure we had good execution and a good rollout.”

McCollum said the company used the second AEDC grant to prepare for expansion and to chase an ambitious goal: “live and accurate data on every single road in the country.

“We have gotten to a really happy point where we know we can handle scaling the company to the size we have ambitions for,” McCollum said. “We want to empower our decision-makers, local, state and federal governments, by providing them with real-word performance analytics.”

How It Works

McCollum calls it Real-Time Pavement Performance Analytics (RPPA), which analyzes roads through sensors attached to the client’s vehicles, such as water department or road department trucks.

The sensors collect data through vibrations while the vehicles carry their passengers on their daily duties. At the end of each route, the sensors send the data recorded to the RMT cloud database.

The RMT platform uses machine learning to analyze the data and determine the conditions of the roads traveled that day. If a road is determined to be in need of maintenance, then the platform sends a notification to the client.

The technology is proprietary and developed by Bates, who earned dual bachelor’s degrees in software engineering and computer science from Auburn before earning a master’s in computer science from Georgia Tech.

“I would just describe him as a boy genius software engineer,” McCollum said of Bates, who is the company’s chief technology officer.

McCollum said RMT meets with prospective clients and analyzes their network of vehicles to devise the best arrangement of sensor carriers. McCollum didn’t give details of that process but said a small city or county could be accurately analyzed by 20 vehicles.

“Regardless of the intention, even if a municipality was going to have a third-party firm go drive every single road every single day and provide them with an evaluation, no human can ingest that much information to make accurate insight,” McCollum said. “At that scale it is completely cost prohibitive. To perform a survey like that even on a monthly basis is completely cost prohibitive.

“Our system educates itself on a daily basis, and it automatically keeps an eye on every road in your network for you. The system is watching the deterioration of every road for them.”

As an example, McCollum pulled up on his screen a section of a city covered by the RMT sensors. Selecting one street showed a dashboard of all the pertinent information, including length, overall quality and repair history.

The grades are broken down by segments on each road. McCollum said this particular road had an A grade for 42% of its length, 33% was a B and 16% was a C. The breakdown also showed how it ranked in the city’s road network.

“We are showing them everything, and this updates every single day,” said McCollum, clarifying that every single day meant every day a vehicle with sensors was driven. “Now when they go into their projects they can say show us every road in the network that is a candidate for [maintenance].”

Ounce of Prevention

Van Buren County Judge Dale James is a believer.

James, who took office in 2019, was one of RMT’s first clients. The county pays $5,000 per year for the system, which can vary in price depending on the size of the sensor fleet. The city of Cabot pays $18,000 per year.

Van Buren County may have only 15,000 citizens but it has 1,100 miles of roads. Because of the modest population, the county’s budget for road maintenance is about $2.5 million, James said, with half of that for payroll with another quarter for equipment.

That doesn’t leave much for actual repairs, which James said can be as much as $100,000 a mile. Knowing where to go before a road becomes a major repair is a great help, just as preventive medicine can help avoid major health problems.

“We are a small county with a limited budget, so it is helping me create an overall better system of roads instead of just listening to the chronic complainers whose roads are never going to be good enough,” James said. “We are operating on a very, very small margin that we can get work done. It is deploying the resources at the right place at the right time to get the most work done. It increases overall efficiency. It cuts out the redundancy of a lot of our work.”

There is an added benefit, James said. The county applies for state and federal grants to help with road repair; Van Buren County has already received two grants because James was able to explicitly show the need.

“The first year I got to use this data, I acquired enough grant funds that we paid enough for 10-plus years,” James said. “When we are all applying for competitive grants, it is easy to say I want this. What we are able to say … because of the unfiltered data is that we need this.”

McCollum said that because of machine learning, no amount of data is too overwhelming.

“What is so valuable about this system is we are the only truly scalable model for collecting this type of accurate data,” McCollum said. “Our next step is at the state level, which we anticipate happening relatively soon.”