‘Why didn't I follow the navigation on this road?’... Kakao Navi learns 'inconvenient roads' with AI

[Financial News] After analyzing inconvenient sections such as long traffic jams, complex alleyways, and congested areas that existing navigation systems could not reflect using artificial intelligence (AI), the functionality has been greatly improved. Kakao Mobility announced on the 14th that it has developed a technology that analyzes driver behavior data on the routes guided by navigation using AI and reflects it in route guidance, proving its effectiveness by applying it to Kakao Navi. The related research paper was published in the top-tier SCI journal 'TRC' in the field of transportation earlier this month.
Kakao Mobility's AI research and development team and Professor Kim Dong-kyu's research team at Seoul National University jointly authored a paper on AI route guidance technology based on driver response, which reflects the potential characteristics of roads that existing navigation systems could not consider in route search based on driver behavior data and confirmed its effectiveness in actual commercial service.
Navigation generally searches for routes based on observable physical information such as 'vehicle speed', 'road width', and 'number of lanes'. However, drivers often deviate from the routes provided by navigation for convenience. This is due to various reasons such as areas with a lot of illegal parking, inconvenient access, or distrust of unfamiliar routes. However, it is practically difficult to reflect all these factors in route guidance.
Kakao Mobility found a solution by comparing and analyzing the routes guided by navigation and actual driving data. In other words, it evaluates the 'traffic value' of roads based on the 'route compliance rate' of whether the driver actually drove on the guided road and reflects it in route search. 
Through this, the system can automatically learn the inconvenient factors that affect the driver's route choice and continuously improve usability without building additional infrastructure. Moreover, it was possible to more precisely calculate the traffic value of millions of road sections nationwide and improve the accuracy and reliability of route guidance by also reflecting real-time traffic information.
In fact, it has become possible to detect inconvenient sections that existing navigation could not reflect, such as roads where users frequently deviate from navigation routes due to queues, complex alleyways, congested areas near transfer centers, and mountainous roads with large elevation differences, and reflect them in real-time route search.
Kakao Mobility has been applying AI route guidance technology based on driver response to Kakao Navi since November last year. When a driver selects a destination, the Kakao Navi algorithm applies this technology centered on 'fast routes', 'highway priority routes', and 'main road priority routes' to propose them as 'navigation recommended routes'.
The effect of this technology was also published in the paper. According to the paper, the analysis of data from the first week of technology application showed that the driver's route compliance rate for newly provided routes increased from 64.22% to 70.87% in 'fast routes', a 6.65%p increase. Similarly, in 'highway priority routes', it increased from 71.32% to 72.91%, and in 'main road priority routes', it increased from 70.79% to 72.40%.
Kim Pureumoe, the first author of the paper and a researcher at Kakao Mobility's AI research and development team, said, "It is academically and service-wise meaningful as we confirmed the improved effect in various route quality indicators such as actual travel time to the destination and road driving convenience."



yjjoe@fnnews.com Yoonju Cho Reporter