[Test Drive] Hyundai to bring futuristic mobility with self-driving tech
Posted : 2017-03-20 16:53
Updated : 2017-03-20 18:52
Hyundai Motor engineer Kim Jin-hak speaks during an interview with The Korea Times aboard the autonomous Ioniq hybrid sedan. The model was autonomously driven without Kim's input during the interview. / Courtesy of Hyundai Motor
By Jhoo Dong-chan
Self-driving vehicles are one of the three most promising technological developments of the automotive industry along with eco-friendly and connected vehicles.
Global top carmakers as well as IT giants such as Google and Cisco have been making big investments to introduce autonomously driving vehicles on the street, and the nation's business bellwether Hyundai Motor has also joined the race.
Hyundai offered The Korea Times an opportunity to board the self-driving Ioniq hybrid sedan at Hyundai Motor Group's Namyang R&D Center in Hwaseong, Gyeonggi Province.
Hyundai Motor engineer Kim Jin-hak was in the driver's seat. The car's self-driving technologies suggested a good grasp of future mobility.
Working together with the vehicle's navigation system, the self-driving mode kicked in by activating cruise control.
The autonomous Ioniq sedan utilized 10 sensors in total, with two cameras mounted on the top of the rear-view mirror and rear side accompanied by two radars and six sensors around the car.
The two cameras monitored other cars and objects as well as traffic lights. They also identified pedestrians, motorcycles and bicycles. A lane departure warning was issued when the vehicle switched lanes without direction.
Inside the grille, front and rear bumpers, radars and sensors provided the vehicle full 360-degree motion detection coverage, Kim said.
"Hyundai Motor has outsourced motion algorithm chipsets for the advanced driver assistance system (ADAS) camera sensors from Mobileye, the world's largest ADAS chipset provider, but started to develop its own motion-detection algorithm in 2010," he said.
"The autonomous Ioniq sedan is now being developed only by Hyundai Motor Group's technologies, completely independent from foreign technologies."
During this year's Consumer Electronics Show (CES) in Las Vegas, Hyundai Motor Vice Chairman Chung Eui-sun showcased the company's self-driving technologies with the autonomous Ioniq sedan. When he rode in the vehicle during the showcase, he took his hands off the steering wheel while drinking coffee and reading a book.
The autonomous Ioniq sedan that Chung tested at the show met the level 4 standard by Society of Automotive Engineers (SAE). SAE level 4 means high automation and level 5 means full automation.
On the streets at the R&D center, the autonomous Ioniq detected approaching vehicles' projected movements, maneuvering the vehicle's headway as safely as it could to prevent collisions.
There was also a seven-inch monitor mounted on the top of the dash showing detected objects and their movements to the driver. It also displayed road signs ahead and possible obstacles like pedestrians and motorcycles.
While traveling a preprogrammed route inside the R&D center compound, the vehicle ran smoothly and uneventfully at speeds between 20 and 45 kilometers per hour. The reporter and Kim could even do an interview while riding in the autonomous Ioniq sedan.
"I believe self-driving technology will be introduced on the nation's highway roads within five years," Kim said.
"Without pedestrians as well as other possible objects, driving on the highway is comparably simple. I believe related technologies are almost there. When it comes to in-city self-driving technology, there are many steps for us to go through. Not only global top carmakers like Hyundai but also IT firms including Google are developing such technologies."
In May, a Tesla driver in Florida was killed after colliding with a crossing trailer while using the vehicle's autopilot mode. Shortly after the accident, Tesla's chipset supplier Mobileye issued a statement that the technology isn't there yet to be considered fully autopilot until 2018, and ended its partnership with Tesla in July.
"There are a number of variables on the street conditions that may influence the self-driving vehicle's safety. When the sensors fail to detect lanes or signals due to foul weather, for example, the vehicle would not be able to autonomously evade accidents as a human driver could do," Kim said.
"In order to minimize such variability, Hyundai Motor is developing the so-called deep learning in its self-driving technologies."
Deep learning is a technology that enables a vehicle to memorize and study patterns of moving objects on the street to more effectively avoid accidents. AlphaGo, an artificial intelligence (AI) of Google, chose its moves based on previously experienced cases "learned" by machine learning to beat human champion Lee Se-dol in a five-game go match last March..
"If an autonomous vehicle algorithm understands and studies such patterns in traffic, like AlphaGo did to beat Lee, it will be a game changer," Kim said.
"I believe it will take more than 10 years to apply a deep-learning algorithm in automotive technologies."