The Effects of Differences in Drivers' Brain Activity on Recognition of Left- and Right-Hand Curves and Driving Performance

Reon Ogawa･Kohjiro Hashimoto･Kikunori Shinohara (Suwa University of Science)･Masashi Makita (Teikyo University)･Hiroshi Kuniyuki (Suwa University of Science)

In this study, the relationship between drivers’ brain activity and driving performance on left- and right-hand curves was analyzed using a driving simulator. The results showed that activation in the right prefrontal cortex was associated with reduced deviation from the curve for both left- and right-hand curves. Furthermore, for right-hand curves, higher activity in the right frontal pole was associated with reduced deviation from the curve.

Analysis of Factors Leading to Disregarding Traffic Signal Accidents

Hiroshi Kuniyuki･Syunya Oshima･Ginpei Suzuki (Suwa University of Science)

In this study, high-accident locations in urban areas where accidents caused by drivers' disregarding traffic signals frequently were investigated and analyzed to clarify contributing factors. The results revealed that these locations are characterized by short intervals between traffic signals at intersections and the presence of a large intersection further ahead. It is considered that short intervals between traffic signals make drivers more likely to perceive these intersections as a single unit, while the presence of a major intersection ahead causes them to overlook the preceding intersection, which are the primary contributing factors to these accidents.

Development of Attribute-Based Driver Models Using a Driving Simulator
-(First Report): Development and application of driver models for cornering-

Shingo Sugiura (TOYOTA Systems)

This paper proposes a method to build attribute-specific driver models using a driving simulator. Using such models is expected to streamline fully virtual ECU evaluation when combined with various test scenarios. We collected simulator data from 41 drivers on a multi-curve course. From the data, we extracted driver features by attribute and built models for younger and older driver groups. We tuned a virtual ADAS using these models and evaluated driver acceptance for the target groups. The results support the effectiveness of the proposed attribute-based driver modeling method.

Evaluation of Stochastic Characteristics of Driver Pedal Operation for Interpreting Machine Learning-Based Pedal Operation Prediction Errors

Akihiro Matsumoto･Toshiya Hirose (Shibaura Institute of Technology)

Toward pedal misapplication detection, machine learning-based prediction of pedal operation has been investigated. However, drivers’ pedal operations are not fully deterministic and may vary even under the same driving conditions. Therefore, when evaluating a prediction model, it is necessary to consider not only the model’s predictive performance but also the stochastic variability inherent in human operation. This study quantitatively evaluates variations in pedal operation under identical conditions and examines basic indices for interpreting prediction errors in machine learning-based pedal operation prediction.