Deep Learning-based Technology for Self-driving Cars
Hyundai Mobis has disclosed its plan to complete the development of artificial intelligence (AI) technology for image recognition within this year and load it onto autonomous vehicles that are scheduled for mass production in 2022.
Hyundai Mobis held a meeting at the 2019 Seoul Motor Show underway at KINTEX, Ilsan, Gyeonggi Province on March 28. In the meeting, the company said that it would secure deep learning-based high performance image recognition technology that can recognize vehicles, pedestrians, and road features and apply it to the sensor of a front camera to support autonomous driving.
This technology will be developed based on image recognition AI which learns image data by automation technique. Hyundai Mobis explained that this technology would enable it to secure most of software and hardware technology applied to camera sensors for autonomous driving.
Hyundai Mobis has also decided to strengthen its collaboration with Hyundai Motor to expand its deep-running technology in the camera field. Through joint development with the carmaker, Hyundai Mobis plans to expand the application of the deep learning-based image recognition technology to a 360-degree surround view monitor (SVM) as well as a front camera sensor for autonomous navigation.
Furthermore, the company aims to take the lead in technology competitiveness in the autonomous driving field by enhancing sensors’ performance through data fusion (sensor fusion) between cameras and radars by combining the image recognition technology with a radar already developed.
In the future, Hyundai Mobis will continue to collaborate with information and communication technology (ICT) companies in the field of deep learning and promote open innovation such as technological cooperation with innovation-powered start-ups and research and development based on industry-academy collaboration. In February, Hyundai Mobis announced a plan to invest about four trillion won in order to sharpen its competitiveness in future car technology for the next three years.