Sho Takahashi |
Faculty of Engineering Civil Engineering Advanced Social System |
Associate Professor |
In the present study, objective is to reveal the driver's subjective risk perception when closing to a preceding vehicle in car-following situation under snowy or dry/wet road conditions. It was assumed that the driver uses the ACC to follow a preceding vehicle, and the preceding vehicle decelerated and approached. The results of the field experiment shown that the driver's subjective risk perception to the preceding vehicle was increased by ACC use. The models for estimating the driver's subjective risk perception to the preceding vehicle based on values of the 1/TTC and the 1/THW in the situation of preceding vehicle deceleration were delivered in each type of road conditions (Risk Feeling equation). From the Risk Feeling equation, the THW of the ACC, which corresponds to the conventional driving of the driver's subjective risk perception to the preceding vehicle, is shown in each road conditions. It was suggested that longer value of the THW is required to reduce driver's risk feeling when driving with ACC on snowy roads.
This study proposed speed adjustment delineators embedded in the center of the travel lane as a preliminary speed adjustment device for drivers who is operating a semi-automated vehicle to reduce conflicts with the merging vehicles on the expressway. We conducted experiments using a driving simulator with 46 participants, and the results of driving records and subjective evaluations showed that the information provided about speed adjustment delineators affected the driver's avoidance behavior. The participants who were explained the intention of the speed adjustment delineators selected the appropriate speed adjustment to follow the instructions of the speed guide light at the upstream of the nose end in the merging section. This result led to merging behavior smoothly and reduced the subjective risk perceived by the drivers due to the merging vehicles. In addition, based on the acceleration and deceleration behavior of drivers when the speed guide lights were not turned on, the proper location of the speed adjustment delineators were revealed.
In winter, to drive a vehicle has a lot of difficulties due to road slipperiness, low visibility and narrowed road by lying snow. Understanding the driving environment where the driver feels danger is necessary to introduce SAE Level 2 or 3 in winter, so the present study aims to clarify driver's risk avoidance behavior when the drivers are using adaptive cruise control (ACC) under the winter road conditions. In an experiment on public road, a total 6 participants drove the test vehicle with ACC-ON conditions on the expressway and several local highways. We measured driver's risk avoidance behavior, and simultaneously measured road slipperiness, road geometry and weather condition. Results indicated that road slipperiness, road geometry and the vehicle speed have an effect on the occurring driver's risk avoidance behavior. Slippery road, hard geometry and high speed driving often lead the driver's risk avoidance behavior. According to this result, it is clarified that driver's risk avoidance behavior was reduced under low regulatory speed conditions on the expressway which the road geometry design was high level whereas the road surface conditions were slippery. It suggests that driver's risk avoidance behavior can reduce if the driving support system slows down before the vehicle approach such a slippery or hard geometry condition road.
Many road videos obtained by CCTV and dashcams are accumulated in organization of road administrators. However, since these data are simply accumulated as original video files, the data transmission via networks is very high cost and huge sized storage is required. Also, in the road situations where some obstacles such as damaged parts of roads and any unexpected objects exist, bicycles cannot travel normally and are forced to avoid them. Therefore, by realizing detection of the avoidance behavior on bicycle trips, more effective road management by road administrators can be expected. In this paper, we propose an edge computing system for accumulating the data of the avoidance behavior on bicycle trips while reducing the data size. This edge computing system contributes to realizing more effective and efficient road management.
The details of the matches of soccer can be estimated from visual and audio sequences, and they correspond to the occurrence of important scenes. Therefore, the use of these sequences is suitable for important scene detection. In this paper, a new multimodal method for important scene detection from visual and audio sequences in far-view soccer videos based on a single deep neural architecture is presented. A unique point of our method is that multiple classifiers can be realized by a single deep neural architecture that includes a Convolutional Neural Network-based feature extractor and a Support Vector Machine-based classifier. This approach provides a solution to the problem of not being able to simultaneously optimize different multiple deep neural architectures from a small amount of training data. Then we monitor confidence measures output from this architecture for the multimodal data and enable their integration to obtain the final classification result.
A novel method for player importance prediction from a player network using gaze positions estimated by Long Short-Term Memory (LSTM) in soccer videos is presented in this paper. By newly using an estimation model of gaze positions trained by gaze tracking data of experienced persons, it is expected that the importance of each player can be predicted. First, we generate a player network by utilizing the estimated gaze positions and first-arrival regions representing players' connections, e.g., passes between players. The gaze positions are estimated by LSTM that is newly trained from the gaze tracking data of experienced persons. Second, the proposed method predicts the importance of each player by applying the Hypertext Induced Topic Selection (HITS) algorithm to the constructed network. Consequently, prediction of the importance of each player based on soccer tactic knowledge of experienced persons can be realized without constantly obtaining gaze tracking data.