Search Results

An investigation based on human visual search functions was conducted into the causative factors of traffic accidents at clear intersections in rural areas. The results indicated that it is difficult for drivers to detect a vehicle traveling on a collision course because the vehicle remains in the same position in the driver's visual field. Two systems are introduced to assist drivers' visual searches. One system uses an image processing technique, and the other utilizes DGPS and IVC techniques. This paper presents the development of the assistance system.

In this paper, an intersection collision warning system using DGPS (Differential Global Positioning System) will be proposed. The system is developed to prevent collisions of vehicles crossing at intersections, especially at well visibility intersections without traffic lights. Two GPS receivers are installed on two vehicles on the move towards the same intersection from different directions. The position and velocity information of the vehicles are measured by on-board GPS receivers, and then transmitted from one vehicle to another by inter-vehicle communication (IVC). Therefore, the relative position and direction of each vehicle and collision judgment coefficient (CJC) which is defined by using the relative position of vehicles are calculated. After taking the crossing position of directions and the variation of CJC into account, the position of the intersection and the possibility of collision can be predicted in advance. Warning will be given to drivers with a prearranged timing.

It was found that the finger blood pulse shows various fluctuations in different driving conditions. The nature of the finger blood pulse fluctuations was used for estimating a driver's internal state. Indexes suitable for expressing the fluctuations were moment and density; these indexes were calculated by using a return-map. However these results were measured by an off-line system and were calculated after the experiment. So, an on-line (real-time) system was needed in order to construct a driver's internal state monitoring system. As a first step, an online system for estimating the human internal state was developed. This system is available for estimating the human internal state every 30 seconds.