Nearing midnight, Dipen Babariya and his companion were traveling to their shared location. The pair utilized Google Maps to find their way but were unsure of where they were. The “quickest path” on the map turned out to be the slowest. Without any street lights, the road was covered in potholes, making it hazardous and impossible to traverse.
“The maps only show the fastest and shortest routes in terms of travel time; they do not show difficulties with the roads, such as poor road conditions. This occurred while we were still in college, which inspired Dipen Babariya to develop an AI-based solution to the issue, according to The Better India.
A college hunch has grown into Bengaluru’s RoadMetrics, a full-fledged firm that offers AI-based data for road management strategies. More than 50,000 kilometers of roads in the nation, including those in Delhi, Bangalore, Mumbai, Jamshedpur, Patna, and others, have had data collected on them.
making Indian roadways more secure and free of potholes
“Mishal Jariwala and I thought the concept had a tonne of promise. We were pursuing engineering at the time with a focus on AI while still in college. The 26-year-old explains that their first idea for 2018 was to create a smartphone application that would alert users of the state of the road.
As Bengaluru is home to hundreds of businesses, Dipen and his co-founder, who are both from Surat, decided to go there to try their hand at entrepreneurship. “It was here that we met our third co-founder, who shared our enthusiasm for the concept. “Nikhil Prasad Maroli had just relocated from the US, where he was employed by automakers like Tesla,” the speaker claims.
The firm drew a lot of interest from investors within the first two months of its formation in 2019 and obtained its initial funding from Sanjay Mehta’s 100X.VC.
“I was a mix of excitement and apprehension. Although I had worked on projects in college, this would be my first experience with entrepreneurship. Road safety is a crucial issue in which no software indicates the condition of the road. India and even nations outside of India can attest to this, adds Dipen.
The three started working on their initial plan to create software for mobile devices. Although sensor-based technology that would function as a mobile application sounded promising, it was impractical because we would need to drive on every inch of the road to record vibrations. The potholes would then be determined by analyzing the vibrations, the expert claims.
Hence, after doing more investigation, Dipen and his co-founders discovered an even more clever approach.
“We created image-based software or computer vision-based. It makes use of a mobile camera that is mounted to a car’s windscreen. After mounting the device, we launch our data collection application. Together with GPS information, locations, and time stamps, it also saves video data. This is uploaded to our servers, which also contain millions of other similar pieces of data. The road defect is then identified by our trained AI program,” he claims.
Our AI technology can recognize ten different sorts of road defects, from minor flaws like surface deterioration and cracks to serious issues like potholes, says Dipen.
In addition to the AI system, the business has created RoadMetrics Maps, which assist users in locating the quickest, coziest, and least congested routes. But, after realizing that the issue required more than simply avoiding potholes, but also filling them, the business decided to transition from a B2C to a B2B strategy.
“Our servers also get the data gathered from the maps. Only Bengaluru has access to the maps for now, he says, adding, “But since this was incredibly labor- and resource-intensive, we chose to go another approach. We assist governments and private players in locating and fixing road issues through the use of a B2B software model. When they can be repaired, why avoid potholes?
While they are working in Assam and Bihar, the business has already completely mapped the cities of Mumbai and Bengaluru.
“The towns and private firms can determine how much funding allocation is necessary based on damage reports evaluated by our program. We are assisting the TATA Group in Jamshedpur in identifying the problems on the roads, he says.
A manual survey takes roughly four to five months to complete, plus another few months for planning.
“What occurs is that the entire state of the road changes by the time the survey is finished. Plans may not be appropriate for the newly discovered damages. During a week, our technology can survey a distance of one thousand kilometers. As a result, it takes less time and is more thorough. I can provide the example of M P Road in Jamshedpur, which would consistently be in poor shape even after repair, as an illustration of that. It turned out that there was a tiny creek close by, and the water was washing the road down below.
“We discovered this issue, and the contractor was able to arrange the road’s repair accordingly,” he continues.
Similarly to this, the business has worked with the Electronic City Township Authority (ELCITA) in Bengaluru to repair the potholes.
