Some drivers' habits shouldn't be replicated.
Hyundai's advanced in-house tech has helped elevate the brand to premium levels, pioneering electric vehicles like the Kona EV and alternative fuels like hydrogen with the Hyundai Nexo. But the carmaker has also been developing new safety tech and has now unveiled the world's first AI-based adaptive cruise control system. Officially, they call it Machine Learning based Smart Cruise Control, or SCC-ML, an industry-first technology that will use artificial intelligence in conjunction with its Advanced Driver Assistance System to learn how you drive and implement your driving patterns into a vehicle's self-driving behavior.
According to Woongjun Jang, vice president of the Hyundai Motor Group, "The new SCC-ML improves upon the intelligence of the previous ADAS technology to dramatically improve the practicality of semi-autonomous features."
Standard Smart Cruise Control already makes use of semi-autonomous functionality, reading traffic flow and speed, and slowing down or speeding up accordingly. Modern iterations of the technology have even been tuned to work in stop-start traffic jams, but many of these systems are irritating as they seem to respond too soon or too late, causing many drivers to simply take over - sometimes too slowly - and do it in a manner that best suits their driving style.
That's where SCC-ML comes into play. The AI systems learn drivers' patterns and habits behind the wheel under everyday conditions, creating a unique profile of the driver. The system then implements this profile into the Smart Cruise Control system to create an autonomous driving style that Hyundai claims is identical to how the driver would handle those situations. Hyundai gives examples of a driver accelerating differently depending on whether a situation is a high-, mid-, or low-speed environment. With the SCC-ML system imitating the driver, it's supposed to give the driver more confidence that the system will react exactly as they would themselves.
Hyundai claims the system can differentiate more than 10,000 patterns, which it can combine innumerably to create an accurate replication of an individual driver's style. The SCC-ML system achieves Level 2.5 self-driving status - Tesla's Autopilot software only achieves Level 2 autonomy status, and its 'Full Self-Driving Capability' is extremely misleading.
This isn't the first time Hyundai has toyed with artificial intelligence, but we can think of a few problems with AI-based adaptive cruise control that learns the drivers' habits, though. The main issue is that not all driving habits are good ones. Many drivers follow far too close to the vehicle ahead of them, not allowing enough space to slow down in the event of an emergency. Erratic acceleration and deceleration is another problem, with drivers regularly applying the brakes too early, too late, or too inconsistently. While the SCC-ML system might make drivers feel more comfortable, it could also pick up on their bad habits, creating more hazards on the road.
The technology is still in development, however, and while it is planned for rollout in future Hyundai and Kia vehicles, it won't happen just yet. Perhaps there's still time to fix the potential flaws, then.