Analysis and Prevention of Road Accidents

Abstract: Road and traffic accidents aren’t Certain and predictable incidents and their analysis requires the knowledge of the factors leading to them. Road and traffic accidents are defined by a set of constraints which are mostly of discrete by nature. The major problem in the analysis of accident data is in it heterogeneous nature. Hence it is to be considered during analysis of the data otherwise, There might be relationship between the data present in the data set may remain hidden.

Introduction

Road accidents analysis represents a valid instrument for all those studies that aim to define road safety. Certain research on different methods pointed out that many countries all over the world have analyzed accidents and traffic data with the purpose of find suitable techniques to identify accident prone locations. Most current methods contemplate the use of statistical procedures for the black spot identification and accident analysis. Although, researchers used segmentation of the data to reduce this heterogeneity using some measures such as expert knowledge, but there is no guarantee that this will lead to an optimal segmentation which consists of homogeneous groups of road accidents.

The usual modern techniques foresee the use of three sets of factors of the highway system that contribute to cause the accidents. These sets of factors are The driver, The Vehicle,The Road environment.To better evaluate the causes of an accident, the present study carefully analyzed the contribution of the last set evidencing the need to divide it into two subsets:

1. Road factors.
2. Environment factors.

Literature Survey

Pierre-Marie Damon, Hicham Hadj-Abdelkader, Hichem Arioui, and Kamal Youcef-Toumi [1] In this paper lateral position of a powered Motor Cycles on the road, its steering behavior and predict the road curvature ahead of the motorcycle. We will be using mapping algorithm for detecting curved or Straight Lanes road. A clothoid model is used to extract pertinent information from the detected road markers.The performance of the proposed approach is illustrated through simulations carried out with the well-known motorcycle simulator “BikeSim.” The results are very promising since the algorithm is capable of estimating, in real time, the road geometry and the vehicle location with a better accuracy than the one given by the commercial GPS. results discussed in this letter are very promising and open lot of possibilities about the use of vision in ARAS developments.

Five steps to obtain the lane fitting from a captured frame. G¨ultekin G¨und¨uz and C¸ a?gdas¸ Yaman and Ali Ufuk Peker and Tankut Acarman [2] In this paper the risk level correlation and legal speed exceeding and average speed ensuing with the human being who is controlling the technical system, i.e. the car, is presented. A dataset is constituted by time stamped and geographically referenced driving maneuver information, which is exceptionally reported when an acceleration exceeds the given threshold in both longitudinal and lateral direction. Risk Level is estimated for amount of Damage that has happen to vehicle when the collision has taken place in that place of the collision that taken on the highway. In this we will also able to predict the future prediction of risk on that place.we elaborate the risk level assessment based on the driving activities exceeding a given threshold for acceleration severity in the longitudinal direction, left and right steering maneuvering severity in the lateral direction, legal road speed exceeding event and average trip speed information.

Zhidan Liu, Zhenjiang Li, Kaishun Wu, and Mo Li [3] In this paper Traffic information is of great importance for urban cities, and accurate prediction of urban traffics has been pursued for many years. These sophisticated models are used to capture hidden traffic characteristics from substantial historical mobility data and then makes use of trained models to predict traffic conditions in the future. We use the deep learning in predictions of various traffic indicators, for example, traffic speed, traffic flow, and accident risk. In addition, we summarize and analyze some early attempts that have achieved notable performance. we envision the potential of rich mobility data and deep learning on urban traffic prediction and discuss some pioneering attempts. Deep learning will advance traffic predictions through powerful representation learning and has shown initial successes we propose two future research directions to improve the accuracy and efficiency of urban traffic prediction on a large scale.

Zhe Peng, Shang Gao, Zecheng Li, Bin Xiao, and Yi Qian [4] In this paper the Information about vehicle safety, such as the driving safety status and the road safety index, is of great importance to protect humans and support them for safe driving. Despite some research on driving safety analysis, the accuracy of driving safety assessment is very limited. Also, the problem of precisely and dynamically predicting road safety throughout a city has not been sufficiently studied. we first discuss mobile sensing data collection in VANETs and then identify two main challenges in vehicle safety analysis in VANETs, i.e., driving safety analysis and road safety analysis. In this we review and classify the state-of-the art vehicle safety analysis techniques into different categories. The results demonstrate the advantages of our proposed scheme over other methods by utilizing mobile sensing data collected in VANETs. The preliminary results demonstrate the effectiveness of our proposed scheme by utilizing mobile sensing data collected.

Dajun Wang , Xin Pei, Li Li , Fellow, IEEE, and Danya Yao, Member, IEEE [5] In this paper Risky driving is a major cause of traffic accidents. We propose a new method that recognizes risky driving behaviors purely based on vehicle speed time series. This method first retrieves the important distribution pattern of the sampled positive speed-change (value and duration) tuples for individual drivers within different speed ranges. Since speed measurement is available on most of the newly build vehicles, this method can be easily implemented and used. This is useful to many traffic applications, e.g., driver training and insurance pricing. To prevent risk driving, we adopt naturalistic driving method to study the driving patterns of individual drivers. So, we can still use similar method to identify risky drivers based on additional GPS position data and growingly richer geographic information system (GIS) information.

Naive Bayes is a simple but an effective classification technique and it is based on the Bayes Theorem. It assumes independence among predictors, that is the attributes or features should be not correlated to one another in anyway, be related to each other. Even if there is dependency, still all these features independently contribute to the probability and that is why it is called Naïve.

1.  

K-means clustering is an unsupervised machine learning algorithm which aims to group similar objects into several cohesive clusters.After the algorithm brings together, new examples can be classified as normal or abnormal by determining which cluster they belong to. The k-means algorithm consists of iterations with more than two steps. [6] First, every example is “colored” assigned to a cluster based on the minimum distance to all cluster centroids. Second, each cluster centroid is updated to the mean value of all the examples in the cluster.

Conclusion

In this study, the technique of association identify the reasons rules with a large set of accident’s data to of road accidents were used.Analysis showed that producing the association rules, makes identification of factors involved in the accident that occur together,easier.It shares a lot in understanding the circumstances and causes of the accident. So the association rule mining gives the direction to deeper research on the causes of road accidents. It helps government to adapt the traffic safety policies with different types of accident and situations. The main result of this study is that although the characteristics of humanity and behavior are very important in occurrence of all road accidents but we can understand that spatial features and infrastructure play a major role in the accident.In this study it is tried to choose the interesting and superior rules to provide a lot of valuable information for policies to provide better safety policies. This article can be a step towards providing useful information for highway engineers and transportation designers to design safer roads.

References

1. [1] Image-Based Lateral Position, Steering Behavior Estimation, and Road Curvature Prediction for Motorcycles Pierre-Marie Damon, Hicham Hadj-Abdelkader , Hichem Arioui, and Kamal Youcef-Toumi (2018)
2. [2] Prediction of Risk Generated by Different Driving Patterns and theirConflict Redistribution G¨ultekin G¨und¨uz and C¸ a?gdas¸ Yaman and Ali Ufuk Peker and Tankut Acarman (2017)
3. [3] Urban Traffic Prediction from Mobility Data Using Deep Learning Zhidan Liu, Zhenjiang Li, Kaishun Wu, and Mo Li (2018)
4. [4] Vehicle Safety Improvement through Deep Learning and Mobile Sensing Zhe Peng, Shang Gao, Zecheng Li, Bin Xiao, and Yi Qian (2018).
5. [5] Risky Driver Recognition Based on Vehicle
6. Speed Time SeriesDajun Wang , Xin Pei, Li Li Fellow, IEEE, and Danya Yao, Member, IEEE(2017).
7. [6] R. Haynes et al., “The influence of road curvature on fatal crashesin New Zealand,” Accident Anal. Prevent., vol. 40, no. 3, pp. 843–850,(2008).
8. [7] L. Bai, Y. Wang, and M. Fairhurst, “Multiple condensation filters for roaddetection and tracking,” Pattern Anal. Appl., vol. 13, no. 3, pp. 251–262,Aug.( 2010).
9. [8] H. Jung, J. Min, and J. Kim, “An efficient lane detection algorithm for lane departure detection,” in Proc. IEEE IV, pp. 976–981(2013).
10. [9] R. M. Faizal and H. M. A. H. Mansor, “Virtual mid-line detectionon curve road for user guidance using simulation model,” in Proc. IEEE ICCTD, 2009, vol. 1, pp. 24–27(2009).
11. [10] L. Bai and Y. Wang, “Road tracking using particle filters with partitionsampling and auxiliary variables,” Comput. Vis. Image Understand.,vol. 115, no. 10, pp. 1463–1471, Oct. (2011).
12. [11] Rasmussen, “Grouping dominant orientations for ill-structured road following,” in Proc. IEEE CVPR, vol. 1, pp. I-470–I-477(2004).
13. [12] J. W. Park, J. W. Lee, and K. Y. Jhang, “A lane-curve detection based on an LCF,” Pattern Recognit. Lett., vol. 24, no. 14, pp. 2301–2313,Oct. (2003).
14. [13] J. Kim, N. Jeon, and H. Lee, “Development of optimization algorithm for green drive on roads with continuous curves,” in Proc. IEEE 38th Annu.Conf. IECON, pp. 2238–2244(2012).
15. [14] J. P. Tarel and F. Guichard, “Combined dynamic tracking and recognitionof curves with application to road detection,” in Proc. IEEE Int. Conf.Image Process.,(2000)
16. [15] Y. Jiang, F. Gao, and G. Xu, “Computer vision-based multiple-lane detection on straightroad and in a curve,” in Proc. IEEE IASP,pp. 114–117(2010).
17. [16] Seibert, M. Hahnel, A. Tewes, and R. Rojas, “Camera based detection and classification of soft shoulders, curbs and guardrails,” in Proc. IEEE IV, pp. 853–858(2013).
18. [17] S. Sayanan and M. M. Trivedi, “Integrated lane and vehicle detection,localization, and tracking: A synergistic approach,” IEEE Intell. Transp.Syst., vol. 14, no. 2, pp. 1–12, Jun. (2013).
19. [18] E. L. Harbeck, A. I. Glendon, ""How reinforcement sensitivity and perceived risk influence young drivers' reported engagement in risky driving behaviors,"" Accident Analysis & Prevention, vol. 54, pp. 73-80, (2013).
20. [19] X. Chen, L. Li, Y. Zhang, ""A Markov model for headway/spacing distribution of road traffic,"" IEEE Transactions on Intelligent Transportation Systems, vol. 11, no. 4, pp. 773-785,(2010).
21. [20] L. Li, X. Chen, Z. Li, ""Asymmetric stochastic Tau Theory in car-following,"" Transportation Research Part F: Psychology and Behaviour, vol. 18, no. 5, pp. 21-33, (2013).
22. [21] L. Li, X. Chen, ""Vehicle headway modeling and its inferences in macroscopic/microscopic traffic flow theory: A survey,"" Transportation Research Part C: Emerging Technologies, vol. 76, pp. 170-188, (2017).
23. [22] A. K. Jain, ""Data clustering: 50 years beyond K-means,"" Pattern Recognition Letters, vol. 31, pp. 651-666, (2010).

Cite this page