Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : September 25, 2019 - September 27, 2019
In recent years, the number of traffic accidents has decreased due to the development of driving support systems. Driving assistance system supports to the driver, however when driving assistance is not appropriate, the driver may be bothered by the assistance and distrusted of the system. In order to provide appropriate support, a method that driver behavior prediction has been proposed. In this research, we measured seat pressure distribution and driving information using a driving simulator, and predicted driving behavior using a convolutional neural network. In this research, we predicted five driving behaviors: "Go straight ahead", "Turn right", "Turn left", "right lane change", and "left lane change”. As a result, it was possible to predict with an accuracy of 65.4% in inter participant evaluation and 96.4% in overall evaluation.
