2:30 PM - 2:45 PM
[HTT21-04] Basic research on predicting the location of human-to-vehicle accidents
Keywords:HUMAN AND VEHICLE ACCIDENT, COMPACT CITY, HAZARD ASSESSMENT
1. Research background/purpose
The number of traffic accidents is decreasing year by year, but the number is still large, and the number of traffic accidents in the 3rd year of Reiwa was 30,5425 [1], which is still a big problem in Japan.
In recent years, research has been progressing not only on environmental and economic aspects but also on environmental and economic aspects of traffic safety, such as Morimoto et al. [2] By comparing the traffic safety of compact cities, it was shown that compact cities tend to reduce traffic accidents. In compact cities, the urban structure is concentrated in a specific area, so it is expected that the distance traveled will be short and that walking and public transportation will be the main means of transportation.
It is expected that there will be more means of transportation on foot from current levels, which, of course, could lead to more human and car accidents. Therefore, the purpose of this study is to predict where pedestrian violations will occur in compact cities.
For predicting the occurrence of a person-to-vehicle accident, the walking speed of pedestrians (specifically, the speed of jumping out when a crossing violation occurs, the walking speed when ignoring a signal, etc.) and the walking position also contribute to the probability of occurrence. Therefore, it is necessary to obtain these data, but since it is necessary to take a realistic method when performing the analysis, the method of performing the analysis from the recorded video is adopted. Use motion analysis software from the video to acquire the trajectory and walking speed of the pedestrian.
3. Past research
According to Mori et al., The percentage of pedestrians arriving at the pedestrian crossing with the [5] traffic light flashing blue or displayed in red is lower for the elderly and those with mobility restrictions than for the non-elderly. However, while 21.9% of offenders ignore the traffic lights, 69.5% of offenders cross the outside of the pedestrian crossing or just before or after the vehicle. [6]
According to Matsui et al. [8], it is known that if both the cyclist and the vehicle driver have poor visibility due to obstacles such as buildings and cannot recognize the other person, the possibility of a traffic accident at the encounter is extremely high. In addition, Ogawa et al. [9] focused on three factors in the road network data: intersection shape (ratio of 4 limbs and 3 limbs), aggregation status of narrow streets, and number of crossable points (crossing points other than intersections). By creating a virtual road network with different three factors, we are analyzing the effect of road network characteristics on the probability of encountering a traffic accident. As a result, when the distance between the departure point and the destination is small, the probability of encountering a two-way traffic accident on the sidewalk is small, and as the distance between the departure point and the destination increases, a one-way traffic accident on the left side of the roadway It was shown that the probability of encounter is small.
4. Research content
The survey area was Sakurai Station Square (South Exit) in Sakurai City, Nara Prefecture, and the flow of people was recorded at Sakurai Station from October 29th to 31st and November 20th to 21st with a fixed camera. In addition, events were held during these periods. The time zone with the largest number of people in the daytime was recorded for one and a half hours, and the trajectory was extracted from the video. As a result, the density of the locus tended to increase in places where there were corners and obstacles.
5. Summary
In this study, we analyzed the passage position and speed of pedestrians using images. From these results, we would like to predict the location of occurrence.
The number of traffic accidents is decreasing year by year, but the number is still large, and the number of traffic accidents in the 3rd year of Reiwa was 30,5425 [1], which is still a big problem in Japan.
In recent years, research has been progressing not only on environmental and economic aspects but also on environmental and economic aspects of traffic safety, such as Morimoto et al. [2] By comparing the traffic safety of compact cities, it was shown that compact cities tend to reduce traffic accidents. In compact cities, the urban structure is concentrated in a specific area, so it is expected that the distance traveled will be short and that walking and public transportation will be the main means of transportation.
It is expected that there will be more means of transportation on foot from current levels, which, of course, could lead to more human and car accidents. Therefore, the purpose of this study is to predict where pedestrian violations will occur in compact cities.
For predicting the occurrence of a person-to-vehicle accident, the walking speed of pedestrians (specifically, the speed of jumping out when a crossing violation occurs, the walking speed when ignoring a signal, etc.) and the walking position also contribute to the probability of occurrence. Therefore, it is necessary to obtain these data, but since it is necessary to take a realistic method when performing the analysis, the method of performing the analysis from the recorded video is adopted. Use motion analysis software from the video to acquire the trajectory and walking speed of the pedestrian.
3. Past research
According to Mori et al., The percentage of pedestrians arriving at the pedestrian crossing with the [5] traffic light flashing blue or displayed in red is lower for the elderly and those with mobility restrictions than for the non-elderly. However, while 21.9% of offenders ignore the traffic lights, 69.5% of offenders cross the outside of the pedestrian crossing or just before or after the vehicle. [6]
According to Matsui et al. [8], it is known that if both the cyclist and the vehicle driver have poor visibility due to obstacles such as buildings and cannot recognize the other person, the possibility of a traffic accident at the encounter is extremely high. In addition, Ogawa et al. [9] focused on three factors in the road network data: intersection shape (ratio of 4 limbs and 3 limbs), aggregation status of narrow streets, and number of crossable points (crossing points other than intersections). By creating a virtual road network with different three factors, we are analyzing the effect of road network characteristics on the probability of encountering a traffic accident. As a result, when the distance between the departure point and the destination is small, the probability of encountering a two-way traffic accident on the sidewalk is small, and as the distance between the departure point and the destination increases, a one-way traffic accident on the left side of the roadway It was shown that the probability of encounter is small.
4. Research content
The survey area was Sakurai Station Square (South Exit) in Sakurai City, Nara Prefecture, and the flow of people was recorded at Sakurai Station from October 29th to 31st and November 20th to 21st with a fixed camera. In addition, events were held during these periods. The time zone with the largest number of people in the daytime was recorded for one and a half hours, and the trajectory was extracted from the video. As a result, the density of the locus tended to increase in places where there were corners and obstacles.
5. Summary
In this study, we analyzed the passage position and speed of pedestrians using images. From these results, we would like to predict the location of occurrence.