Hyundai Motor Group said Monday that it has developed the world’s first self-driving technology using AI-based machine learning, which detects the driver’s driving pattern and apply it to the smart cruise control system.

In recent years, a host of carmakers have added advanced driver assistance systems to their models for safer and more convenient driving. 

One of the core technologies is adaptive cruise control, a function which makes a car slow down and speed up automatically to keep pace with the car in front.

Until now, a driver was to set the maximum speed for cruise control and only had up to 4 different control options. To maintain the selected speed, a radar sensor watches for traffic ahead, locks the car in a lane and instructs the car to stay 2 to 3 seconds behind the car ahead of it, keeping their distance from those in front.

According to Hyundai Motor, its latest self-driving technology called SCC-ML -- Smart Cruise Control-Machine Learning -- works by consistently gathering driving data via front camera and radar sensors, so the machine learning algorithm can study the driver’s driving pattern, preference and habits.

Hyundai Motor said such technology considers driving preference in largely three categories: distance with a car in front, acceleration and responsiveness to the driving environment, such as traffic conditions and obstacles on the road.

SCC-ML can differentiate some 10,000 different driving patterns on the road, so anyone can experience customized smart cruise control driving and does not feel a sense of difference even when driving on cruise control mode, the company explained. 

“SCC-ML applied AI-based machine learning algorithm to significantly improve the usability of smart cruise control system,” said a Hyundai Motor official, adding that it was “meaningful” for them to have secured the AI-based self-driving technology for the first time in the industry. 

As cruise control is a crucial part of the self-driving cars, Hyundai Motor said it will gradually apply SCC-ML to its new cars. 

The carmaker claimed that SCC-ML can be considered “level 2.5” self-driving technology, as it includes lane change assist and highway driving assist, which provides autonomous steering, acceleration, deceleration and distance keeping mode while cruising on a highway. 

Currently, most cruise control technology offers level 2 autonomous driving.
According to US Department of Transportation’s National Highway Traffic Safety Administration (NHTSA), autonomous driving can be measured on a scale of 0-5: At level 0, all major systems are controlled by humans; at level 1, certain systems such as cruise control or automatic breaking can be controlled by the car.

Level 2 involves at least two simultaneous and automated functions such as acceleration and steering, but human control is required to ensure safety. At level 3, the car is able to self-manage safety functions, but the driver is expected to take over when alerted. At level 4, the car is fully autonomous in most driving scenarios. The final level 5 refers to when the car is completely autonomous.

By Kim Da-sol (ddd@heraldcorp.com)