Search

This study aimed to construct driver models for overtaking behavior using long short-term memory (LSTM). During the overtaking maneuver, an ego vehicle changes lanes to the overtaking lane while paying attention to both the preceding vehicle in the travel lane and the following vehicle in the overtaking lane and returns to the travel lane after overtaking the preceding vehicle in the travel lane. This scenario was segregated into four phases in this study: Car-Following, Lane-Change-1, Overtaking, and Lane-Change-2. In the Car-Following phase, the ego vehicle follows the preceding vehicle in the travel lane. Meanwhile, in the Lane-Change-1 phase, the ego vehicle changes from the travel lane to the overtaking lane. Overtaking is the phase in which the ego vehicle in the overtaking lane overtakes the preceding vehicle in the travel lane. Lane-Change-2 is the phase in which the ego vehicle changes lanes from the overtaking lane to the travel lane after overtaking the preceding vehicle in the travel lane. We analyzed and evaluated the accuracy of the models for the entire experimental scenario and for each model of Lane-Change-1, Overtaking and Lane-Change-2. The constructed models were also evaluated for safety in lane-change situations based on the WP.29 steering equipment regulation of the UNECE, which was defined as Scritical in the Automatically Commanded Steering Function (ACSF) Category C of R79 in WP.29 of UNECE. In the results, the Lane-Change-2 model was identified to be the most accurate of the four phases and the model accuracy was found to be due to epochs and learning rate. All the constructed models satisfied the WP.29 regulation of UNECE. The constructed models which reflected individual characteristics are expected to be applied to ADAS and automated driving technology that conduct overtaking maneuvers.

The purpose of this study was to construct driver models using long short-term memory (LSTM) in car-following situations, where other vehicles change lanes and cut in front of the ego vehicle (EGV). The development of autonomous vehicle systems (AVSs) using personalized driver models based on the individual driving characteristics of drivers is expected to reduce their discomfort with vehicle control systems. The driving characteristics of human drivers must be considered in such AVSs. In this study, we experimentally measured data from the EGV and other vehicles using a driving simulator consisting of a six-axis motion device and turntable. The experimental scenario simulated a traffic congestion scenario on a straight section of a highway, where a cut-in vehicle (CIV) changed lanes from an adjacent lane and entered in between the EGV and preceding vehicle (PRV). To construct a highly accurate model, we analyzed critical variables as input information affecting the output of the LSTM model using a random forest (RF) model. The results showed the high importance of the EGV velocity, THW, and relative velocity as information related to the traveling lane, in addition to the CIV velocity as information related to the CIV. The CIV data obtained after the lane change were used for the PRV in this analysis. Based on the variables analyzed in the RF model, we constructed personalized driver models using LSTM, and the mean coefficient of determination was greater than 0.95, indicating that this system is more accurate than the conventional car-following models. The driver models constructed in this study are expected to improve the usability of AVSs employing the driver model.

For car-following models, the car-following characteristics differ depending on the vehicle type, such as passenger cars, motorcycles, and trucks. Therefore, constructing a model for each category is essential. To that end, various modeling methods have been proposed; however, herein, we particularly focused on the long short-term memory (LSTM), which is the best method for forecasting long-term time-series data.[1, 2] The objective of this study was to construct a car-following model for each vehicle category using the LSTM and to evaluate the model accuracy for each vehicle category. In this study, US-101 and I-80 data provided by the next-generation simulation (NGSIM), which is based on natural traffic flow data, were used. In the NGSIM, only car-following situations were selected as car-following data, and these were classified into the vehicle type: motorcycles, passenger cars, and trucks. The classified data were then used to construct LSTM-based car-following models for each category. As the evaluation of this model was dependent on the amount of data, the model was built according to the amount of data for motorbikes, which was the lowest. A merged integrated model was also constructed by mixing the data obtained from all vehicle type, and the accuracy of the model was compared with that of the model for each vehicle category. Thus, this study investigated the impact of car-following models constructed for each vehicle category on the model accuracy using the LSTM. The results showed that for passenger cars and motorcycles, the RMSE values of the vehicle-specific models were smaller than those of the integrated model and the accuracy of the models was better than that of the integrated model. In future, we intend to construct car-following models for each vehicle characteristic.

This study clarified the features of fatal truck-pedestrian and truck-cyclist accidents by a comparison with sedans. Fatal accident data from the Japanese Institute Traffic Accident Research and Data Analysis (ITARDA) database were analyzed for the years 2012 ∼ 2016. The trucks were classified into three categories: heavy-, medium-, and light-duty. All three categories had higher percentages of pedestrians as fatal collision partners as opposed to cyclists. Every truck category had significantiy higher percentages of pedestrian fatalities due to daytime right turns than sedans. Heavy-duty trucks had significantly higher percentages of cyclist fatalities due to daytime and nighttime left turns than sedans. These results show that the design specifications for new sensing technology to avoid collisions with pedestrians and cyclists should differ among vehicle types. The knowledge obtained in this study will be useful to develop new technology or systems that are effective in dangerous situations between trucks and pedestrians or cyclists.

Vehicles that start moving from a stationary position can cause fatal traffic accidents involving pedestrians. Ultrasonic sensors installed in the vehicle front are an active technology designed to alert drivers to the presence of stationary objects such as rigid walls in front of their vehicles. However, the ability of such sensors to detect humans has not yet been established. Therefore, this study aims to ascertain whether these sensor systems can successfully detect humans. First, we conducted experiments using four vehicles equipped with ultrasonic sensor systems for vehicle-forward moving-off maneuvers and investigated the detection distances between the vehicles and a pipe (1 m long and having a diameter of 75 mm), child, adult female, or adult male. The detections of human volunteers were evaluated under two different conditions: front-facing and sidefacing toward the front of each vehicle. Front-facing is defined as the condition where the human faces the vehicle front, while side-facing is that where the side of the human faces it. For both the front-facing and side-facing conditions, the results indicated that the sensor-detection distances for a child were shorter than those for the pipe, whereas those for adults were less than or approximately equivalent to those for the pipe. These results revealed that ultrasonic sensor systems for vehicle-forward movingoff maneuvers can detect not only stationary objects but also humans, indicating that ultrasonic sensors installed in the vehicle front could possibly reduce the risk of vehicle-forward moving-off accidents involving pedestrians.

The factors in fatal injuries to cyclists in collisions with five different types of vehicles were examined using data from the Institute for Traffic Accident Research and Data Analysis in Japan for the period 2004 ~ 2013. The vehicle types were sedans, minivans, box vans, light passenger cars (LPCs), and light cargo vans (LCVs). The analyses focused on cyclist age, gender, main body regions that suffered fatal injuries, and vehicle travel speeds. Cyclists aged 75+ years made up the largest group of fatalities in collisions with sedans, box vans, LPCs, and LCVs, whereas those aged 13 ~ 64 years made up the largest group of fatalities in collisions with minivans. Head injuries were the most frequent injuries in cyclist fatalities. The percentage of head injuries in fatalities of cyclists aged 75+ years was significantly higher than that for cyclists aged 13 ~ 64 years for box vans, but significantly lower for sedans, LPCs, and LCVs. For sedans, the percentage of hip injuries in fatalities of cyclists aged 75+ years was significantly higher than that of cyclists aged 13 ~ 64 years. Cyclist fatalities due to hip injuries were significantly more common for females than for males in collisions with sedans, LPCs, and LCVs.

We investigated a driver’s gazing time during backing vehicle maneuvers in a 90-degree reverse parking system for the phase from start to the 5-m point (start phase) and the phase from backing from the 5-m point to stop (stop phase). The objective of this study was to elucidate the driver’s gaze while approaching the parking area and in the parking area during backing. We used a sport utility vehicle with the driver’s seat at the right-hand side. We modified the vehicle with four technological conditions: no-tech vehicle, sonar vehicle, backup-camera vehicle, and backup camera with sonar vehicle. The monitor, rearview mirror, left-side mirror, and right-side mirror were designated as the area of interest (AOI) in the gaze. For the start phase, there were no significant differences in the total gazing time of all the AOIs in the four technologies. For the stop phase, however, the total gazing time of all the AOIs for the vehicle with a backup camera was significantly longer than that of the vehicle without a backup camera owing to the increase in the gazing time of the monitor. The results obtained from these experiments will form a basis for the future development of vehicle camera systems.

Ultrasonic parking sensors are an active technology designed to alert drivers to the presence of objects behind their vehicle but not the presence of a human. The purpose of this study was therefore to ascertain if these sensor systems can successfully detect a human subject. We accordingly conducted experiments using four vehicles equipped with both rear-facing center and corner ultrasonic parking sensor systems to determine the detection distance between the vehicle and a 1-m tall, 75-mm diameter pipe, a child, an adult woman, and an adult man. The detection of human subjects was evaluated under front-facing and side-facing conditions behind each vehicle. The results indicate that for a front-facing and side-facing child, the center sensor detection distances were 50-84% and 32-64%, respectively, shorter than that of the pipe. For front-facing and side-facing adults, the center sensor detection distances were just less than or roughly equivalent to that of the pipe at 89-102% and 78-97%, respectively. A similar trend was seen for the corner sensors. Notably, under the side-facing condition, the sensor detection distances were slightly shorter for all subjects than under the front-facing condition. These results reveal that ultrasonic parking sensor systems can not only detect objects but also humans, indicating that ultrasonic sensors are an available countermeasure to prevent backover accidents involving pedestrians. To address the shorter detection distance of children, a combination of ultrasonic parking sensors with other systems, such as backup cameras, may be more effective for avoiding backover collisions.

To assess the effect of wearing a bicycle helmet using an adult headform in terms of the head injury criterion (HIC) when the frontal and lateral parts of the helmet impact a vehicle body and also when the frontal part of the helmet impacts the road surface.The adult headform was made to impact the hood, windscreen, roof top, and roof side rail of a vehicle body at an impact velocity of 35 km/h, which is a common head-to-vehicle impact velocity in real-world cyclist-vehicle collisions, in which the vehicle impacts the cyclist at 40 km/h. For the road surface impact experiments, we set a drop height of 1.5 m (impact velocity of 20 km/h).Helmet usage helped to reduce the HIC when the frontal and lateral parts of the helmet impacted vehicle parts other than the hood. The HIC reduction for the frontal impact was greater than that for the lateral impact. Moreover, the higher the stiffness index of the vehicle structure, the greater was the HIC reduction. However, helmet usage was ineffective for reducing skull fracture risk (HIC 2558) when the lateral part of the helmet impacted stiffer parts of the vehicle, such as the roof side rail close to the B-pillar. Helmet usage helped to reduce the HIC by 91% when the frontal part of the helmet impacted the road surface.Wearing a helmet reduces skull fracture risk when the frontal and lateral parts of the helmet impact vehicle parts (excluding the hood) at 35 km/h and the road surface at 20 km/h. However, when the lateral part of the helmet impacts the B-pillar, the helmet cannot effectively reduce the skull fracture risk at these real-world velocities.

The Japan New Car Assessment Program (J-NCAP) evaluates the performance of cars in terms of protection against whiplash injuries in rear-end collisions. In the test protocol, a simplified triangular acceleration is applied to the sled. This study clarifies whether biofidelic rear-impact dummy II (BioRID II) measurements obtained for simplified triangular acceleration reflect car-to-car rear-end impacts in real-world accidents in Japan. We conducted a car-to-car rear-end impact experiment and a simplified-triangular-acceleration sled test. Our results indicate that the time series of dummy responses were approximately consistent in the two test conditions. The neck injury criterion (NIC) and maximum acceleration of the head and T1 measured using the BioRID II dummy were similar in the car-to-car and sled experiments. This revealed that the J-NCAP test protocol using simplified triangular acceleration reflects the car-to-car rear-end impact experiment using Japanese cars, in terms of the NIC and maximum acceleration of the head and T1.

The incidence of traffic accidents in Japan has been decreasing annually. Nineteen percent of all accidents involve bicycles, with 51 % of these accidents being at road crossing intersections. Therefore, to reduce the number of accidents, this study analyses driving and cycling characteristics and proposes suitable collisions prevention methods. First, the study measured traffic environment variables using video cameras at a target non-signalized intersection and analyzed the speed and time to intersection of bicycles and motor vehicles. Thus, 47 dangerous situations were observed via the video analysis, and most of these situations occurred when the vehicle’s time to intersection ranged from 0.50 to 0.75 s and the bicycle’s speed ranged from 2.0–3.0 m/s. Second, using the results of video camera analysis as experimental parameters (e.g., the speed and timing of the presence of the bicycle), this study conducted an experiment with a driving simulator to investigate the effect of warning drivers about the risk of collision. A driver support system was then utilized to provide acoustic and optical warnings to drivers. The experiments revealed that the motor vehicle time to the anticipated collision point (V-TTC) increased with the use of a driver support system. Significant differences between experiments with and without driver support systems were observed when the calculated time between the bicycle and the motor vehicle was 0.25 and 0.50 s. Therefore, when the calculated time was 0.25 and 0.50 s, a driver support system, indicating the presence of a bicycle, was effective in preventing an intersection collision.

Objective: This study aimed to investigate the situational characteristics of fatal pedestrian accidents involving vehicles traveling at low speeds in Japan. We focused on vehicles with 4 or more wheels. Such characteristics included daytime or nighttime conditions, road type, vehicle behaviors preceding the accident, and vehicle impact locations. Methods: Pedestrian fatality data on vehicle-pedestrian accidents were obtained from the Institute for Traffic Accident Research and Data Analysis of Japan (ITARDA) from 2005 to 2014. Nine vehicle classifications were considered: Trucks with gross vehicle weight (GVW) ≥7.5 tons and

This study clarifies vehicle accident scenarios in which pedestrians suffer serious injuries. We investigated the characteristics of serious pedestrian injuries in accidents, using vehicle-to-pedestrian accident data for 1999 and 2009 in Japan. We focused on the main body region of seriously injured pedestrians considering different factors including the accident year, vehicle type, travel speed, and pedestrian gender and age. The results indicate that sedans, mini vans and light passenger cars tended to injure pedestrians’ legs more frequently, while box vans and light cargo vans tended to injure the head more frequently. The results also show that female pedestrians suffered hip injuries more often than male pedestrians in accidents involving sedans, mini vans, light passenger cars and light cargo vans. In accidents involving sedans, it was revealed that the head and hip were injured statistically more frequently at high vehicle travel speed than at low vehicle travel speed in all age groups.

Vehicle occupants were killed in 33% of all fatal traffic accidents in Japan in 2014. The abdomen is associated with the highest risks of serious injuries and fatalities, so we have investigated the causes of abdominal injuries to vehicle occupants. Fifty-four percent of vehicles struck in vehicle-versus-vehicle accidents were impacted from the front. An investigation on accident details determined that lap belts caused serious abdominal injuries. Residual deformation marks suggest that seats move vertically during a frontal impact. The purpose of this study was to clarify the effect of vertical movements of the seat on the lap belt in the event of a frontal impact. We investigated the phenomenon using simulations and an anthropomorphic 50th percentile male Hybrid III dummy model wearing a three-point seat belt. Our results revealed that the lap belt slipped from its initial position around the dummy waist when the seat height was lowered by 150 mm. We concluded that abdominal injuries could be ca...

The effect of eye movement on the gaze area was investigated using a driving simulator when a driver simply followed a preceding vehicle on a straight road under two conditions; while focusing on a vehicle and while focusing on a vehicle and red points in a wide area of the front screen. The area of the distribution of the gaze points was greater in the latter case, especially on the right-hand side. The results suggest that a tool that forces a driver to gaze across a wide area would encourage gazing in side areas. The study also investigated the cognition time of a driver as other vehicles suddenly emerged at intersections. The cognition time at the first intersection was significantly longer than that at the last intersection, suggesting that the first event of an approaching car enlarged the area of the driver's vision as the driver displayed greater caution.

The vehicle travel velocity at the moment of pedestrian contact is considered to be an important parameter that affects the accident outcome. To reduce the vehicle-to-pedestrian travel velocity, a pedestrian detection system using a sensor in a vehicle could be an effective countermeasure. The specifications of the pedestrian detection system will need to be designed properly based on real-world accident situations. Then, the purpose of this study is to clarify the relation between vehicle travel velocity and pedestrian injury severity or fatality risk in traffic accident analyses in order to propose specifications of a pedestrian detection system. We analyzed the relation using the Institute for Traffic Accident Research and Data Analysis (ITARDA) database for 2009 in Japan, considering different factors including a vehicle type and pedestrian age in actual traffic accidents. The results indicate that the percentage risk of fatality was consistent between sedans and light passenger cars at every vehicle travel velocity, rising sharply at a vehicle travel velocity of 40 km/h. Meanwhile, in car-to-pedestrian accidents involving sedans and light passenger cars, pedestrians aged 60 and over faced the higher risk of fatality, which increased rapidly at the vehicle travel velocity of 40 km/h. Therefore, the present study proposes that the specifications of the pedestrian detection system should be designed so that the vehicle travel velocity in traffic accidents could be reduced to 30 km/h or lower, when we consider the decrease of fatality.

Accident emergency calling systems (AECSs) are signaled by the deployment of airbags, which causes them to automatically emit information providing the location of the accident site to a public service answering party (PSAP). In some realworld accidents, airbags have failed to deploy. This study clarifies the factors that influence the nondeployment of front airbags in vehicle-vehicle collisions, investigating nondeployment of the driver-side front airbags in sedans and light passenger cars (LPCs) from Japanese accident data. The component rates of deployment for front airbags tend to be higher than those of nondeployment at higher values of pseudo-ΔV in vehicle-vehicle frontal impacts. For both sedans and LPCs, the transition zones between nondeployment and deployment of the front airbag occur at pseudo-ΔV values of 30-50 km/h (ΔV ≈ 21-35 km/h). For mutual impact locations where sedans and LPCs impact opponent vehicles at pseudo-ΔV ≥ 40 km/h (ΔV ≈ 28 km/h) in frontal impacts, the component rate of front airbag nondeployment is higher than that of deployment in right-to-right impacts. The results indicate that factors influencing front airbag nondeployment in vehicle-vehicle collisions are ΔV, impact offset configuration, and crossing angle. Considering front airbag nondeployment in real-world accidents, AECSs should have other functions, such as a manual button, to emit information in addition to automatic emission via airbag signaling.

This study aimed to clarify the relationship between truck-cyclist collision impact velocity and the serious-injury and fatality risks to cyclists, and to investigate the effects of road type and driving scenario on the frequency of cyclist fatalities due to collisions with vehicles. We used micro and macro truck-cyclist collision data from the Japanese Institute for Traffic Accident Research and Data Analysis (ITARDA) database. We classified vehicle type into five categories: heavy-duty trucks (gross vehicle weight [GVW] ≥11 × 103 kg [11 tons (t)], medium-duty trucks (5 × 103 kg [5 t] ≤ GVW < 11 × 103 kg [11 t]), light-duty trucks (GVW

To reduce the severity of injuries and the number of cyclist deaths in traffic accidents, active safety devices providing cyclist detection are considered to be effective countermeasures. The features of car-to-bicycle collisions need to be known in detail to develop such safety devices.The study investigated near-miss situations captured by drive recorders installed in passenger cars. Because similarities in the approach patterns between near-miss incidents and real-world fatal cyclist accidents in Japan were confirmed, we analyzed the 229 near-miss incident data via video capturing bicycles crossing the road in front of forward-moving cars. Using a video frame captured by a drive recorder, the time to collision (TTC) was calculated from the car's velocity and the distance between the car and bicycle at the moment when the bicycle initially appeared.The average TTC in the cases where bicycles emerged from behind obstructions was shorter than that in the cases where drivers had unobstructed views of the bicycles. In comparing the TTC of car-to-bicycle near-miss incidents to the previously obtained results of car-to-pedestrian near-miss incidents, it was determined that the average TTC in car-to-bicycle near-miss incidents was significantly longer than that in car-to-pedestrian near-miss incidents.When considering the TTC in the test protocol of evaluation for safety performance of active safety devices, we propose individual TTCs for evaluation of cyclist and pedestrian detections, respectively. In the test protocols, the following 2 scenarios should be employed: bicycle emerging from behind an unobstructed view and bicycle emerging from behind obstructions.

This study aims to clarify the relation between axonal deformation and the onset of axonal injury. Firstly, to examine the influence of strain rate on the threshold for axonal injury, cultured neurons were subjected to 12 types of stretching (strains were 0.10, 0.15, and 0.20 and strain rates were 10, 30, 50, and 70 s-1). The formation of axonal swellings and bulbs increased significantly at strain rates of 50 and 30 s-1 with strains of 0.15 and 0.20, respectively, even though those formations did not depend on strain rates in cultures exposed to a strain of 0.10. Then, to examine the influence of the strain along an axon on axonal injury, swellings were measured at every axonal angle in the stretching direction. The axons that were parallel to stretching direction were injured the most. Finally, we proposed an experimental model that subjected an axon to more accurate strain. This model observed the process of axonal injury formation by detecting the same neuron before and after stretching. These results suggest that the strain-rate dependency of axonal tolerance is induced by a higher magnitude of loading strain and an experiment focusing on axonal strain is required for obtaining more detailed injury criteria for an axon.

The purpose of this study was to identify and better understand the features of fatal injuries in cyclists aged 75 years and over involved in collisions with either hood- or van-type vehicles.This study investigated the fatal injuries of cyclists aged 75 years old and over by analyzing accident data. We focused on the body regions to which the fatal injury occurred using vehicle-bicycle accident data from the Institute for Traffic Accident Research and Data Analysis (ITARDA) in Japan. Using data from 2009 to 2013, we examined the frequency of fatally injured body region by gender, age, and actual vehicle travel speed. We investigated any significant differences in distributions of fatal injuries by body region for cyclists aged 75 years and over using chi-square tests to compare with cyclists in other age groups. We also investigated the cause of fatal head injuries, such as impact with a road surface or vehicle.The results indicated that head injuries were the most common cause of fatalities among the study group. At low vehicle travel speeds for both hood- and van-type vehicles, fatalities were most likely to be the result of head impacts against the road surface. The percentage of fatalities following hip injuries was significantly higher for cyclists aged 75 years and over than for those aged 65-74 or 13-59 in impacts with hood-type vehicles. It was also higher for women than men in the over-75 age group in impacts with these vehicles.For cyclists aged 75 years and over, wearing a helmet may be helpful to prevent head injuries in vehicle-to-cyclist accidents. It may also be helpful to introduce some safety measures to prevent hip injuries, given the higher level of fatalities following hip injury among all cyclists aged 75 and over, particularly women.

The vehicle travel velocity at pedestrian contact is considered to be an important parameter that affects the crash outcome. To reduce vehicle/pedestrian impact velocity, a collision damage mitigation braking system (CDMBS) using a sensor for pedestrian protection could be an effective countermeasure. The first purpose of this study is to clarify the relation between vehicle travel velocity and pedestrian injury severity due to differences in pedestrians’ ages in actual traffic accidents. The accident analyses were performed using vehicle-pedestrian accident data in 2009 from the database of the Institute for Traffic Accident Research and Data Analysis (ITARDA) in Japan. The result revealed that the fatality risk became higher with the increase in vehicle travel velocity. The second purpose of this study is to determine the safety performance of production vehicles equipped with the CDMBS for pedestrian protection. It was found that the CDMBS was highly effective in reducing the impact velocity from 50 km/h (vehicle travel velocity) to below 17 km/h, that could result in a significant decrease in fatality risk to be 2% or less. Additionally, the authors investigated a detectable zone with respect to a pedestrian’s position in relation to the vehicle. It was shown that the detectable zones for production vehicles tested were limited to be inside the vehicle front width.

The number of traffic deaths in Japan was 4,612 in 2011. Looking at the road accident fatalities, it revealed that pedestrians accounted for the highest number in 2011 (1,686, 36.6%). To develop safety countermeasures to decrease the severity of injuries and to reduce the number of deaths in traffic accidents, the detailed characteristics of pedestrian injury in vehicle-to-pedestrian crashes are necessary. The purpose of this study is to understand the scenarios of vehicle accidents in which pedestrians suffer fatal injuries. In the present study, we investigated the characteristics of pedestrian injuries in fatal crashes from accident analyses and compared them to head injury severity levels in impact tests against a road pavement and vehicle contact surfaces. In the accident analyses, we investigated the main body regions injured, that is, the most serious extent of injuries over the whole body of pedestrians by using macro vehicle-pedestrian accident data from database of the Institute for Traffic Accident Research and Data Analysis (ITARDA) of Japan. In comparing the differences in injury frequencies for various body regions between 1999 and 2009, it is noted that the frequencies of pedestrian fatalities due to head injuries were significantly reduced. The result indicated that even though head injuries were the most frequent cause of pedestrian fatalities in traffic accidents, the introduction of pedestrian head protection regulation in Japan in 2005 could be considered effective in reducing fatal head injuries. On the other hand, the frequencies of pedestrian fatalities due to hip injuries increased significantly in sedans, light passenger cars, and light cargo vans. Using the macro accident data in 2009, we investigated the frequency of pedestrian fatalities by gender, age group, vehicle travel speed, and fatal head injuries due to vehicle impacts or road pavement impacts. The results also indicated that the frequencies of female pedestrian fatalities due to hip injuries were significantly higher than those of males. Additionally, the results showed that the frequency of pedestrian fatalities due to hip injuries for the age group of over 65 years was significantly higher than that of the industrial age group (aged 13-59). Focusing on the frequency of pedestrian fatalities due to hip injuries in all age groups, vehicle travel speed appeared likely not to be an extremely important factor in increasing fatal hip injuries. In examining the differences in the frequency of fatal head injuries due to contacts with vehicles or road pavements, it is noted that injury frequency in crashes involving vehicles travelling at high speeds were significantly higher than those at lower speeds for pedestrians over 60 years of age. Focusing on head injury severity levels in impact tests against a road and vehicle surfaces using an adult pedestrian headform impactor, the impacts against a road pavement (HIC 6525) was considered to be more severe than those against vehicle front components with high stiffness (HIC from 2600 to 4032). Language: en

The purpose of this study is to clarify the mechanism of traffic accidents involving cyclists. The focus is on the characteristics of cyclist accidents and scenarios, because the number of traffic accidents involving cyclists in Tokyo is the highest in Japan. First, dangerous situations in traffic incidents were investigated by collecting data from 304 cyclists in one city in Tokyo using a questionnaire survey. The survey indicated that cyclists used their bicycles generally while commuting to work or school in the morning. Second, the study investigated the characteristics of 250 accident situations involving cyclists that happened in the city using real-world bicycle accident data. The results revealed that the traffic accidents occurred at intersections of local streets, where cyclists collided most often with vehicles during commute time in the morning. Third, cyclists' behavior was observed at a local street intersection in the morning in the city using video pictures. In one hour during the morning commute period, 250 bicycles passed through the intersection. The results indicated that one of the reasons for traffic accidents involving cyclists might be the combined effect of low visibility, caused by the presence of box-like building structures close to the intersections, and the cyclists' behavior in terms of their velocity and no confirming safety. It was observed that, on average, bicycle velocity was 3.1 m/s at the initial line of an intersection. The findings from this study could be useful in developing new technologies to improve cyclist safety, such as alert devices for cyclists and vehicle drivers, wireless communication systems between cyclists and vehicle drivers, or advanced vehicles with bicycle detection and collision mitigation systems.

This study aimed to clarify the relationship between truck-pedestrian crash impact velocity and the risks of serious injury and fatality to pedestrians. We used micro and macro truck-pedestrian accident data from the Japanese Institute for Traffic Accident Research and Data Analysis (ITARDA) database. We classified vehicle type into five categories: heavy-duty trucks (gross vehicle weight [GVW] ≥11 × 103 kg [11 tons (t)], medium-duty trucks (5 × 103 kg [5 t] ≤ GVW < 11 × 103 kg [11 t]), light-duty trucks (GVW

Protection of vulnerable road users is important to reduce traffic fatality in Japan. The use of active safety devices that can detect cyclists is considered to be an effective countermeasure for reducing the severity of injuries and number of cyclist fatality. The detailed features of car–cyclist contact scenarios need to be clarified to develop such safety devices. Since there is limited information on real-world accidents, the present study investigates near-miss scenarios captured by drive recorders installed in passenger cars. In this study, we showed the similarities of the contact situations between near-miss car-to-bicycle incidents and real-world fatal cyclist accidents. Then we analyzed video frames of 161 near-miss data when bicycles were crossing the road in front of the forward-moving cars to understand the contact situations between cars and bicycles. Additionally, using a video recorded by a drive recorder, we calculated the estimated collision time from the car velocity and the distance between the car and bicycle. For the definition of the estimated collision time, the worst situation was assumed to be that when a car moved toward a cyclist without the driver realizing the cyclist and braking. The results indicated that the average estimated collision time in the cases where bicycles emerged from behind obstructions was shorter than that in the cases where drivers had unobstructed views of the bicycles. When we consider the estimated collision time in the test protocol of evaluation for safety performance of active safety devices, two scenarios should be employed; bicycle emerged from behind unobstructed views and bicycle emerged from behind obstructions.

The main purpose of this study is to define the relationship between the car impact velocity and serious injury risk or fatality risk of cyclists. The authors investigated the risks of serious injuries and fatalities of cyclists using vehicle-cyclist accident data from the database of the Institute for Traffic Accident Research and Data Analysis (ITARDA) in Japan. The vehicle types considered are sedans, mini vans, box vans, light passenger cars and light cargo vans. The results revealed that a 10-km/h decrease in the impact velocity could reduce the severe injury risk and fatality risk for impact velocities of 40 km/h or higher. Specifically, when the impact velocity was less than or equal to 30 km/h, the serious injury risks were less than 21% and the fatality risks were less than or equal to 1% for the above listed vehicle types. Therefore, if the Collision Damage Mitigation Braking System (CDMBS) equipped vehicles can perform its functions effectively so as to reduce the impact velocities, then cyclist injuries will likely be significantly reduced. Another purpose of this study is to assess the effect of wearing a helmet for protection of the cyclist's head. Impact experiment results showed that the measured head injury criterion (HIC) with helmets are lower than that of head-form impactor without a helmet, reducing the HIC by 57%.

The paper presents a fine-grained walking activity recognition toward an inferring pedestrian intention which is an important topic to predict and avoid a pedestrian's dangerous activity. The fine-grained activity recognition is to distinguish different activities between subtle changes such as walking with different directions. We believe a change of pedestrian's activity is significant to grab a pedestrian intention. However, the task is challenging since a couple of reasons, namely (i) in-vehicle mounted camera is always moving (ii) a pedestrian area is too small to capture a motion and shape features (iii) change of pedestrian activity (e.g. walking straight into turning) has only small feature difference. To tackle these problems, we apply vision-based approach in order to classify pedestrian activities. The dense trajectories (DT) method is employed for high-level recognition to capture a detailed difference. Moreover, we additionally extract detection-based region-of-interest (ROI) for higher performance in fine-grained activity recognition. Here, we evaluated our proposed approach on "self-collected dataset" and "near-miss driving recorder (DR) dataset" by dividing several activities-crossing, walking straight, turning, standing and riding a bicycle. Our proposal achieved 93.7% on the self-collected NTSEL traffic dataset and 77.9% on the near-miss DR dataset.