Evaluation of Effectiveness of Next-generation Lane Change Assist System for Reducing Driver Burden

Shoei Tomikawa･Saki Shindo (Honda Motor)

In recent years, technology for driving assistance functions has advanced, with features developed to suggest lane changes to drivers in scenarios such as encountering a slow vehicle in front or exiting highways. This presentation will explain the content of lane change assistance proposal functions designed to address unique scenes specific to Japan, which have not traditionally been addressed, and their effectiveness.

Prediction of Merging Vehicle Position using a Rule-based Model based on Machine Learning Insights and Actual Vehicle Validation

Yuta Takashima･Taku Umeda･Tomoki Uno･Yuko Omagari･Kazuo Hitosugi (Mitsubishi Electric)

To achieve safe driving in merging areas, it is necessary for mainline vehicles and merging vehicles to predict their movements and coordinate with each other. In this study, we developed a model to predict the merging positions of vehicles using machine learning and leveraged the insights from this to develop a high-accuracy, low computational cost predictive model based on rule-based methods. Furthermore, we verified its effectiveness through real vehicle tests.

Collision Avoidance Assist for Distracted Drivers using Gentle Deceleration

Takuya Niioka･Yugo Yamaguchi･Yoshihiro Oniwa (Honda Motor)

The number of accidents has decreased these days due to the widespread application of autonomous emergency braking system. For further reductions, however, it is necessary to address the distracted drivers, the primary cause of most accidents, and support them to safely avoid crashes. This paper introduces a method to support collision avoidance by prompting early recognition with gentle vehicle behavior, unlike conventional approaches that involve large braking forces.

LKA, a Lane Departure Prevention and a Haptic Information on Risk

Shoma Edamoto･Shuji Kimura･Tsutomu Tamura･Robert Fuchs (JTEKT)

Lane Departure Prevention (LDP) can reduce the occurrence of traffic accidents by correcting the vehicle heading in the event of a lane departure. We propose the LDP based on admittance control in this presentation. The proposed method can maintain the vehicle in the lane strictly when the driver is hands off while providing haptic information to the driver according to the distance to the lane marking when driver is hands on.