2A2-H01 運転支援に向けた合流運転行動のモデリング(【交通・物流部門】カー・ロボティクスとITS)

Abstract

In the present paper, driver models of merging behaviors into the main lances from a side lane was proposed for application to driver assistance systems. The first model, a merging position judgment model determines the position where the driver merges based on the states of vehicles driving in the main lanes before merging. The second model, a merging timing judgment model judges whether the host vehicle can merge into the main lane in the current states of the vehicles. Driving simulator experiments with 10 participants were conducted to build models. The accuracy of the models was evaluated by comparing the drivers' behaviors and the model output. The rate of correct judgment by the merging position judgment model was 79% according to the experiments. Furthermore, a prototype of a merging assistance system was developed using the proposed driver models. The simulator experiments with the assistance system demonstrated that high subjective rating by the participants for the appropriateness of the assistance and decrease of mental workload in merging.