RAISE 2020: Deep tech startup RoadMetrics crowdsources data to map road information at scale
Bengaluru-based deep tech startup RoadMetrics’ AI-based solution uses image and sensor data obtained from smartphones to help enterprise mapping firms and smart city bodies with analytics on road networks.
Two years ago, Dipen Babariya and Mishal Jariwala were on their way to a friend’s house in Surat. Google Maps suggested an alternate time-saving, shorter route. However, poor road conditions and inaccurate information meant it took the duo twice the time to reach their destination.
This pain point led the childhood friends to contemplate creating a solution that provided road and street-level data. The duo worked on road condition mapping software in college for two years and submitted it as a project in the final year.
They won the best project award and were soon contacted by Surat Municipal Corporation for a pilot road assessment for the city to help with road maintenance and town planning.
In 2019, this breakthrough led to the inception of RoadMetrics, a deep tech startup on a mission to map road and street data at scale.
The co-founders moved to Bengaluru to take their vision forward. They attended various tech conferences and startup meets in India’s Silicon Valley. At one such event, they met Nikhil Prasad Maroli, an industrial and operations engineer from Texas A&M University, who had returned from the US after working with Velodyne LiDAR and Tesla.
Together, the trio began working to make their vision a reality.
RoadMetrics is an AI-based solution that uses image and sensor data obtained from a simple smartphone to classify road defects, signs, traffic signals, streetlights, etc. This road and street-level data helps enterprise mapping firms and smart city bodies with analytics on road networks and a better mapping experience.
RoadMetrics was one of the nominees at RAISE 2020. RAISE, or ‘Responsible AI for Social Empowerment’, was a summit organised by NITI Aayog and the Ministry of Electronics and Information Technology.
![[RAISE 2020] This AI-powered startup offers accurate, contextual translations in 22 Indian languages](https://images.yourstory.com/cs/2/98c65090592f11ea9f62339ce853ca75/Image64wg-1602661997080.jpg?fm=png&auto=format&h=100&w=100&crop=entropy&fit=crop)
How does RoadMetrics work?
The startup uses images and smartphone sensor data to classify road defects. The classification is done by machine learning and AI-based algorithms.
Nikhil says, “We use a marketplace platform to acquire crowdsourced data. As soon as it is uploaded to the cloud, our team processes it with our AI-based software that identifies various road/street assets and features. At this point, RoadMetrics has a coverage of over 2,500 km in three cities, including Bangalore, Surat, and Jamshedpur with over 100,000 image points acquired.”
The data is sold through the API model as well as on a per km basis, ranging from Rs 700 to Rs 1,500, depending on the geography for Smart Cities.
RoadMetrics has completed a pilot project for JUSCO (Tata Group) in Jamshedpur and is now in talks with mapping/automotive companies like Mahindra for partnerships.
The primary target customers of RoadMetrics are mapping and automotive companies and smart cities for intelligent alerts and warnings, ADAS systems, and analytics-driven road maintenance. The secondary target audience includes location intelligence solutions for businesses based on road/street information.
Nikhil says RAISE 2020 gave the startup “a platform to showcase itself in the global forum and participate with startups across categories”. “We are delighted to be a winner among 24 startups and receive exposure and guidance from the government.”
Challenges and the competition
Talking about the major challenges, Nikhil says training the software requires thousands of data points that are not readily available for Indian roads. The company has tied up with a cab fleet to acquire this data of over 2,500 km for Bengaluru. He adds that standardising operating procedures, including lighting, R&D, and costing, takes time.
RoadMetrics sees itself as a data company for road and street-level data. With more than 5.9 million km of roads in India, unlocking this data with a crowdsourced marketplace solution is a tremendous opportunity.
While competitors like Sweden-based Mapillary and US-based RoadBotics have been able to penetrate the Western market, Nikhil says no player is doing this at scale in India and the emerging economies.
The way ahead
Last December, RoadMetrics raised a seed funding round from 100X.VC, and three angel rounds with convertible ISAFE notes. The company is now equipped to scale up the marketplace platform.
Nikhil says the company aims to map all Tier I cities of India by July 2021 and then target Tier II and III cities.
The RoadMetrics mobile app is similar to Google Maps/Waze. But the difference is that important road and street data is flanked by the marketplace platform used to crowdsource data so users are compensated for the data they acquire, either through direct payments per km or points that can be redeemed for rewards.
By early 2021, the company plans to enter the HD mapping space with the addition of LiDAR sensors that provide a complete 3D mapping solution with centimetre-level precision.
HD mapping is a useful component for self-driving and ADAS systems as well as a useful data repository for mapping companies for higher precision road/street-level information.
RoadMetrics aims to focus on Bengaluru with the available funding. “Our Series A funding will allow us to expand to other major metros in India with HD mapping,” Nikhil says.
Edited by Teja Lele