This paper describes about two kinds of methods for a car driver to maintain alertness using visual stimuli on a head-up display (HUD). In the first method based on a feedback of driver state, the visual feedback stimuli (FB) moving synchronized with horizontal eye movement, the color of which is determined by arousal level estimated with percentage of eye closure (PERCLOS), are presented on a HUD. In the second method focused on environment outside the car that a driver cannot perceive easily, the attention control stimuli (AC) starts blinking before appearance of a passing vehicle, the color of which changes whether a driver gazes at the vehicle or not. Time courses of PERCLOS, subjective sleepiness and annoyance with the visual stimuli during presentation of 20 minutes driving movie were investigated in conditions with FB, AC and without any stimuli (control). Mean PERCLOS for 10 subjects increased with time and the subjects reported stronger sleepiness in the latter period of the movie in all three conditions. An alertness period of time until the PERCLOS exceeds 10% was 14, 14 and 10 minutes in FB, AC and control condition, respectively. Nine of ten subjects answered least annoyance level in FB and AC condition. These results show that the proposed methods are effective and acceptable to extend alertness.

View full abstract

In this study, we investigated a method of using brain waves to predict driver disaccord under the assumption that the driver is monitoring the behavior of an automatically controlled vehicle. This method is aimed at providing feedback to the vehicle control system as to whether the vehicle is operating according to the driver ’s expectations without any disaccord. A task of monitoring the vehicle ’s automatic execution of a lane change was created using a driving simulator. The brain waves of 23 male participants were recorded while they monitored the vehicle behavior. The results revealed that error-related brain potentials appeared when the participants experienced disaccord. Furthermore, a machine learning algorithm was trained to distinguish between brain waves with and without disaccord, obtaining good accuracy rate in discrimination tests. This result suggests that disaccord information can be fed back to an autonomous driving control system and that a neurofeedback system might possibly be applied to vehicles in the future.

View full abstract

We have implemented the personal mobility performing semiautonomous control by estimating the avoidance direction of pedestrian and the avoidance judgment timing of passenger. In the space of pedestrian and personal mobility coexist, it is possible to realize a safety collision avoidance by moving the mobility in the lateral before passenger's avoidance judgment. Therefore, we utilize a Microsoft Kinect which obtains the coordinates of both feet of pedestrian. And we have tested collision avoidance experiments between a pedestrian and the mobility with the implemented model estimating avoiding direction from the feet's relative positions. We have evaluated the pupil size and the number of saccades of passenger by eye camera. As a result, it is possible to reduce the mental stress by semiautonomous control of mobility than autonomous control. Furthermore, we have evaluated the important body site in avoiding judgment and avoidance judgment timing. Based on the results, we have proposed a model of passenger's vision guidance and mobility avoidance.

View full abstract

In car navigation systems, route guidance messages which can be understood with less cognitive load and without any visual information are required. We study effective guidance methods using a comparison between human guidance and current navigation systems. Experimental results show that the guidance messages including visible landmarks are useful for drivers' understanding. Also, the messages should be provided at the appropriate timing for drivers' understanding of the route and driving behaviors. Based on these results, we pursue guidance systems using environmental objects and driver's status monitoring devices.

View full abstract

Advancement of self-driving technology makes it possible to operate with high precision, such as maintaining a stable vehicle speed, which is difficult for the typical driver, maintaining short distances between vehicles, and reacting very quickly to changing conditions. Although, there are few studies on how the driver feels such automatic control. For example, it is not understood what level of anxiety will be felt by passengers when the vehicle parks next to a neighboring vehicle with only a few centimeters to spare, or drives at a speed that results in discomfort. In the experiment reported in this paper, two of the components (“Distance to other vehicles” and “Speed of vehicle”) were used as variables in a driving test performed using an autonomous vehicle simulator. The level of anxiety and whether or not it was acceptable were recorded for each variable using a questionnaire. A model was created for the relationship between the aforementioned variables and the sense of anxiety felt by individuals, and an “acceptable value” was computed in the form of “A distance (speed) of up to X is acceptable.”.

View full abstract

The human-centered automation principle, saying that the human should have the final authority over the automation, has been regarded as the essential design requirement of automated systems. However, the reliability of human performance can be decreased by the effects of time pressure, high workload, and so on. Therefore, adaptive automation systems, which are characterized as the dynamic function allocation between the human and the automation, are expected. In order to realize such systems, the estimation of operators’ workload are necessary. The present research, therefore, has developed a workload estimation method using the physiological data of an operator. A wearable sensing device called JINS MEME was introduced to obtain operators’ electrooculography (EOG), acceleration, and gyro sensor data while they handled a complex simulation task provided by Smart Grid Simulator. A machine learning method, Support Vector Machine, has successfully identified two types of categories of operators’ workload conditions, “High” and “Acceptable”, over 90% accuracy using 10 parameters based on JINS MEME outputs. In addition, based on the detailed analysis of individual differences including each parameter, the effective utilization method of machine learning in workload estimation for adaptive automation has been discussed.

View full abstract

Recently cars that have functions of autonomous cruising and parking have become popular. However, even if with those functions, drivers have to operate manually at difficult parking places that require human decisions. In this research, we developed and evaluated a display system embedded on a steering wheel that assists drivers ’poor steering action at parking. This paper describes that as a result, we obtained evaluation results that the trial display are able to assist drivers ’poor parking skills.

View full abstract

Digital Fabrication has become familiar to ordinary people, but making data for digital tools such as a Laser cutter using CAD is difficult for beginners. We developed an easy to learn and accurate CAD interface for tablet “Drafter” based on stationary metaphor. We compared Drafter and present CADs by experiments consisting of both drawing part and questionnaire part. The drawing part showed that Drafter is easier to learn than present CADs, has the same accuracy as present CADs. The questionnaire part showed that Drafter has the same accuracy as present CADs, and is better at usability than present CADs.

View full abstract

In this study, we developed two features of the balloons presentation of the universal puppetry system for appreciation by people with hearing-impairment. One of the developed feature is the automatic follow-up feature for balloon dialogues. This is a feature of measuring the body position information of the doll operator and determining the presentation position of the balloons based on the information. Anoter is the balloon presentation feature of expression emotions. This is a feature of changing the size and form of the balloon in order to express the character's emotions. These are expected that possible to reduce the possibility that the shade of the doll obstructs reading of the balloons. We conducted the evaluation experiment with the university student with the hearing impaired. Analysis of the results showed that the automatic follow-up feature can help people with hearingimpaired understand the contents of the balloon, simultaneously recognize it with the behavior of the doll, identified which character is speaking. And these also showed that the balloon presentation feature of expression emotions can help people with hearing-impaired understand the character's emotions.

View full abstract