Lack of parking spaces is one of the biggest problems not only inKorea but also around the world. One of our clients, Happy Robot, is solving this problem with AI and robotics. We introduce how Happy Robot is solving this problem and how it made a leap forward through the Data Voucher Support Project and Datahunt.
Start with the Data Voucher Support Program with Datahunt!
The Datavoucher Support Project is a project that provides government support for data purchase and processing costs in the form of vouchers to early and mid-sized companies, small and medium-sized companies, and budding entrepreneurs. DataHunt has been selected as a supplier for four consecutive years since 2019, and was selected as a best practice in 2020, garnering attention from the data processing industry. Here are some real-life examples of companies that have collaborated with Datahunt.
Building learning data through datavoucher
I'm curious about the background of Happy Robot's participation in the datavoucher
The system that Happy Robot is developing is an embedded AI (EAI). Happy Robot's EAI is designed for parking robots that move cars. It is an innovative way to efficiently use parking spaces by having low-height parking robots lift and move cars, and it is essential to have AI inside the robot. We realized that the parking robot needed 120mm of clearance to operate smoothly. We needed to collect images of the car while sticking to the ground like a real robot's gaze, and collect data for the AI to learn the height of the car. From collecting the data to processing it, it was a very challenging task.
When selecting a vendor for the datavoucher, Happy Robot's first priority was to ensure that the vendor was capable of handling the difficult data collection and processing tasks mentioned above. It was also essential that the vendor understood the EAI thatHappy Robot was developing. After discussing proposals with several vendors, we decided that DataHunt was the most qualified and began working with them.
Tell us about your journey with Datahunt in supporting the data voucher program.
The training data that Happy Robot requested fromDataHunt was to derive a height measurement method for a large number of car models and to take pictures from the robot's point of view. The AI engineers at Datahunt graciously stepped up to the plate to complete this mission.
First, they defined the data to be built based on their understanding of the AI to be developed by Happy Robot. Next, they took pictures on the floor of the parking lot to collect the data. It was a very difficult mission, but they collected high-quality images and accurately annotated them, which was very helpful for AI training. Thanks to their enthusiasm, we were able to successfully complete the AI training that HappyRobot was developing.
Happy Robot's Successful Strategy for Datavoucher
What are your plans and goals for the future collaboration between Happy Robot and Datahunt?
In the future, Happy Robot will continue to collaborate with Datahunt to build relevant data sets, including EAI development. With the successful completion of the data voucher support project, we have promised that we will continue to work together in the future.Happy Robot will create a robot that will change the world, and Data Hunt will be a part of it.I hope that those in charge of demand companies who will read this article will also choose a supplier company like Datahunt that can work together for mutual innovation and growth. You'll find a great partner in the data voucher support business.
Lastly, what are some of the most important things to consider when choosing a vendor for datavoucher?
Under the same goal of AI adoption, the demand side and the supply side are not working separately, they are working together to achieve a common goal. Therefore, I recommend that you work with a supplier who understands the AI model that you are developing, and who is willing to actively and enthusiastically work to solve the problems that you need to solve to actually implement your model. That's a place like Datahunt.