Cruise control have been offered on cars for a long time. Basically, the system maintains a set speed and distance from the car ahead based on the inputs provided by the driver. However, the system had limitations as it could not mimic what the driver would have done in a situation when she was driving. As a result, the cars maintained distances and reacted in various situations at their own pace. This made several drivers jittery and uncomfortable as they would have responded differently and did not feel in control of the vehicle.
Step in machine learning. Machine learning is basically a system wherein the car studies the behaviour of the driver through various sensors like speed sensors, braking sensors, steering inputs etc. Thereafter artificial intelligence algorithms understand the driver behaviour under various conditions and learns how to respond in a manner similar to the driver. For example, a driver may be maintaining shorter distance between cars in a congested and slow city driving situation while at the same time maintaining a larger distance from the car ahead on a motorway. When the car learns these patterns and implements them in cruise control mode, the driver feels more in control. Thus she may keep the cruise control on now, instead of switching it off as she did earlier.
Hyundai Motors has developed a Machine Learning based Smart Cruise Control called the SCC-ML. The system has been developed by integrating artificial intelligence with Advanced Driver Assistance System (ADAS) offered by Hyundai. This smart cruise control will be offered with future Hyundai models. This system creates a customised cruise control for every driver. Smart Cruise Control (SCC) enables an essential self-driving feature and core technology for ADAS; which is ; maintaining distance from the vehicle ahead while traveling at the speed selected by the driver.
The driving pattern of the driver that the SCC-ML learns comprises, distance maintained by the driver from preceding vehicles, the rate at which she accelerates, and responsiveness of the driver, ie, how quickly she responds to driving conditions. In addition the system also monitors driving condition and speed. In all, the system learns 10,000 patterns. With Hyundai’s upcoming Highway Driving Assist system that features automatic lane change assist, SCC-ML achieves Level 2.5 self-driving. The system regularly updates the driving pattern information with sensors, thus reflecting the driver’s latest driving style. Additionally, the SCC-ML is programmed specifically to avoid learning unsafe driving patterns in order to increase its reliability and safety.