About Kyung Hee

Home > About Kyung Hee > Media & Press > News



NO.364 01.15.2018

Engineering Kyung Hee: Autonomous Car “KHUV” Starts!

On December 13, 2017, the KHUV (pronounced kju:b), an autonomous vehicle developed by a research team led by Professor Soon-Geol Lee of the Department of Mechanical Engineering, successfully completed the on-campus driving demonstration at the Global Campus. As one of the key areas of Industry 4.0, the autonomous vehicle segment is expected to make significant breakthroughs in the future along with artificial intelligence (AI) technology.


Direct control is possible without additional external devices
Professor Soon-Geol Lee said, "An autonomous car is a holistic intelligent driving technology that recognizes and judges the surrounding external factors and risks posed without the involvement of human driver, and controls and expresses the vehicle accordingly. At this stage we are currently studying autonomous driving in cars, but our aim is to eventually have an autonomous navigation function for all mobile objects. "


Other autonomous vehicles currently under development use external actuators to control the vehicle. However, if an external actuator is attached, the vehicle will not be certified by the Ministry of Land, Infrastructure and Transport and the driving on the ordinary roads will be limited. The team has developed a way to allow internal control without an external actuator.


With the press of a button, KHUV can be switched between autonomous and normal driving modes, and is ready for street driving. Since there is no external actuator, it is expected that it will be able to receive autonomous driving vehicle certification by the Ministry of Land, Infrastructure and Transport.


At the KHUV autonomous driving demonstration, President Inwon Choue said, "I watched the video clip in which an AI robot, Sophia, was talking with UN Deputy Secretary-General Amina Mohammed on October 11, 2017, at the regular meeting of the UN Economic and Social Council (ECOSOC) and the speed of development of AI is very fast. We need to ensure that this rapid technological development would not be abused or unjustly exploited.”


He then added, “Since autonomous driving is an AI that everyone desires, we are looking forward to rapid development. We would like to express our respect and gratitude to professors and students who have made great achievements in such short amount of time.”



Next Challenge is high-speed driving and response to unpredicted situation
The team applied Deep Learning to KHUV. It allows objects to be recognized without any restriction on speed, so that prediction can be made even when multiple vehicles are overlapped. The team also used SSD (Single Shot MultiBox Detector) algorithm which recognizes objects on the screen as multiple boxes and performs recognition in a unified framework.


This improves recognition accuracy and speeds data processing. The data presented here recognize objects in real time using the You Only Look Once (YOLO) algorithm and perform autonomous driving. The YOLO algorithm is up to 1000 times faster than conventional algorithms.


Professor Soon-Geol Lee said, "The current technology can detect, avoid forward obstacles, accelerate or decelerate without a driver. We plan to develop technologies that allow high-speed driving and accident avoidance. We hope that we will transfer the autonomous vehicle technology developed by us to the industry and become one of the exemplary cases to lead Industry 4.0."

Facebook 0
Twitter 0

Return to News

News List

Highlight News