Risk Factors Associated with Accident Severity In Urban Chennai
Road traffic injuries alone account for nearly 3000 deaths per day and around two million lives worldwide. India, developing nation also faces substantial number of death due to road accidents and it is increasing at an alarming rate. Among several states, Tamil Nadu topped the number of road accidents accounting for around 14.9% of the road accidents occurring in the country. Among the cities, Chennai reported the highest number of road accidents during 2016. Hence systematically analysing road data and implementing proper measures are need of the hour to reduce fatality rate in coming years in urban cities. Hence in this study, road traffic data extracted from the Road accidents analysis and management system (RADMS) database has been analysed for urban Chennai for the year 2015. The outcome variable was considered to be accident severity and the predictor variables were classified into driver, road, vehicle, temporal, collision and environmental factors. The quantitative data was analysed for descriptive statistics and chi- square test were performed to measure the association between accident severity and predictor variables at 95% confidence intervals. Driver’s gender, no of drivers involved in accidents, Whether the accident was hit and run case, location of accident, vehicle type, Illumination level, road surface conditions, central divider, junction type, no of lanes, traffic movement and landmark were the contributing factors associated with accident severity. Data analysis from this study provide insights on factors to be considered to reduce accident fatalities and injuries.
Reference:
- Sathish, K. S. and V. Balasubramanian (2017), Risk Factors Associated With Accident Severity In Urban Chennai, Proceeding of the International Conference on Ergonomics and Human Factors (HWWE, 2017), India, 8-10 December
Analysis of Hit-and-run Crashes in Indian Metropolitan City by Logistic Regression Models
Hit-and-run crashes refer to accidents where drivers leave the spot of accident without helping the victims or reporting the accidents to the concerned authorities. The trend of hit-and-run has been increasing in developing nations like India where recent report by ministry of road transport and highways of India showed that among total accidents that occurred in the nation 11.6% accounted for hit-and-run crash (MoRTH, 2016). Using the data collected from the crash database from Urban Chennai, the present study applied logistic regression to identify the factors associated with hit-and-run crashes. Dataset used in this study consists of 4148 samples among which 670 (16.15%) were identified as hit-and-run crashes. A total of 65 variables under 20 factors were classified based on human factors approach including driver, vehicle, crash and environmental factors. The results of present study indicate that unidentified driver, drivers who possess valid driving license, day light conditions, location, crash type, undivided road and public places are important determinants of hit-and-run crashes. Also, the results showed a good accuracy of 92.3% and ROC area under the curve was 0.924. The results of present study provide important insights to reduce mortality due to hit-and-run crashes in cities in developing nations.
Hit-and-run cases have also seen a steep escalation in the past few years. Recent published report by ministry of road transport and highways shows that 55,942 hit-and-run cases were reported which is 11.6% of the total accidents that occurred in the nation (MoRTH, 2016). In the current study, various factors associated with hit-and-run crashes were investigated for urban Chennai, a metropolitan city of India. A total of 65 variables under 20 factors were classified as driver, vehicle, crash and environmental factors. Data mining technique, classification and regression tree (CART) was used on the data related to 4818 hit-and-run crashes that occurred in Chennai urban between January 2015 and December 2016. The dataset was split in two as training and testing data with 50:50 ratios. The predictive accuracy of the model built with total of 65 variables was 92.29% for the training data and 92.19% for the testing data. The CART findings show that collision type is the most important variable associated with hit-and-run crashes. Other secondary variables associated were gender, driver age, vehicle type and light conditions. From the results of the present study, it can be concluded that CART algorithm can be a useful tool in determining and identifying potential causes of hit-and-run accidents.
Analysis of Driver Injury Severity in Metropolitan Roads of India Through Classification Tree
Reducing the injury severity from traffic accidents is most important step in mitigating accidents occurring in developing economies like India where two-way roads are more common in cities. The number of deaths due to accidents has rose from 83,491 in 2005 to 1,36,071 in 2016 as per the latest reports of ministry of road transport and highways, government of India (MoRTH, 2016). To explore the factors contributing to injury severity in such roads, non-parametric classification tree is used since it does not assume any underlying assumption between target variable and the predictors. CART (Classification and Regression tree), a classification tree establishes empirical relation between injury severity outcomes and variables including driver, vehicle, crash and environmental factors. The present study analysed traffic crash data of single lane two-way roads of Chennai city pertaining to period from January 2015 to December 2016. The final dataset included a total of 5271 crash information after excluding incomplete and missing data. This finalized dataset was split into two subsets, training and testing data and the classification models reported an accuracies of 63.4% and 61.5% for the training and testing data. The results indicated that collision type and vehicle type were the two important variables affecting the severity of injury. The findings of this study will help in determining influential factors so that countermeasures to reduce the severity of injury in urban cities can be developed.
Analysis of Traffic Injury Severity in Motor Cycle Riders – A Case Study from India
Powered Two Wheelers (PTW) provide a flexible, faster mode of transport in congested traffic. This is rapidly growing mode of transportation systems in most of the middle-income countries like India and also in urban metros of developed economies. According to the latest annual report by ministry of road transport and highway in India 2015, the number of registered two wheelers increased by 178% in the past decade i.e., from 47 Million in 2003 to 132 Million in 2013. PTWs safety is of major concern since crash reports show that PTWs contribute to 33.8% of total accidents reported (MoRTH,2016). Data mining technique, classification tree was used on the data related to 3002 motor cycle crashes that occurred in Chennai city, India between January 2015 and December 2016. Injury severity was response variable most sensitive to crashes occurring in the city. The influential factors contributing to crashes were classified into driver, vehicle, crash and environmental factors according to human factors approach. Collision type, Driver age and Season are the top three influential factors in determining the severity of PTW accidents. The accuracy of the model built was 66.2% for the training data and 60.5% for the testing data. The findings of this study might help traffic engineers, road safety researchers to develop targeted countermeasures to reduce PTW accidents.
Analyzing Speeding Behaviour and other Driver Related Errors in Indian Metropolitan City
According to the latest published report from the ministry of road transport and highways (MoRTH,2016), driver errors is the single most important factor responsible for causing road accidents. Driver errors alone have accounted for 84% of the total road accidents occurred in India during 2016 compared to 77%, the previous year. Exceeding unlawful speed alone has accounted for 66.5% of driver error related faults. In the present study, we examine the speeding behaviour of drivers and attempt to identify significant risk factors associated with this traffic violation. Risk factors including driver characteristics, vehicle, environmental and crash/ accident characteristics were considered. We analysed the most recent crash data (N=5543 ) of two years (2015-2016) for the Indian metropolitan city, Chennai from the data source provided by RADMS (Road Accident Data Management System).Contingency tables were constructed to assess the association between the risk factors and speeding behaviour with chi square tests of independence at a level of 5% significance. To further estimate the effect of different predictor variables on the likelihood of occurrence of speeding, logistic regression analyses were carried out. This study identified that male drivers, without valid driving license, single lane uncontrolled junction roads involving all type of vehicles in presence of median separators during daylight have higher probability of speeding related crashes. The result from the present study helps in enforcing rules to reduce speeding rate in a country with highest number of accidents.
