A data-driven on-site injury severity assessment model for car-to-electric-bicycle collisions based on positional relationship and random forest

Ye Yu; Zhiyuan Liu; Ye Yu; Zhiyuan Liu

doi:10.3934/era.2023173

Electronic Research Archive

2023, Volume 31, Issue 6: 3417-3434. doi: 10.3934/era.2023173

Previous Article Next Article

Research article Special Issues

A data-driven on-site injury severity assessment model for car-to-electric-bicycle collisions based on positional relationship and random forest

Ye Yu ¹,
Zhiyuan Liu ^{2
,
,}

1.
Department of Public Security Management, Jiangsu Police Institute, Nanjing, Jiangsu, China
2.
Jiangsu Key Laboratory of Urban ITS, Jiangsu Province Collaborative Innovation Center of Modern Urban Traffic Technologies, School of Transportation, Southeast University, Nanjing, China

Received: 08 March 2023 Revised: 05 April 2023 Accepted: 11 April 2023 Published: 18 April 2023

Vulnerable road users (VRUs) are usually more susceptible to fatal injuries. Accurate and rapid assessment of VRU injury severity at the accident scene can provide timely support for decision-making in emergency response. However, evaluating VRU injury severity at the accident scene usually requires medical knowledge and medical devices. Few studies have explored the possibility of using on-site positional relationship to assess injury severity, which could provide a new perspective for on-site transportation professionals to assess accident severity. This study proposes a data-driven on-site injury severity assessment model for car-to-electric-bicycle accidents based on the relationship between the final resting positions of the car, electric bicycle and cyclist at the accident scene. Random forest is employed to learn the accident features from the at-scene positional relationship among accident participants, by which injury severity of the cyclist is assessed. Conditional permutation importance, which can account for correlation among predictor variables, is adopted to reflect the importance of predictor variables more accurately. The proposed model is demonstrated using simulated car-to-electric-bicycle collision data. The results show that the proposed model has good performance in terms of overall accuracy and is balanced in recognizing both fatal and non-fatal accidents. Model performance under partial information confirms that the position information of the electric bicycle is more important than the position information of the cyclist in assessing injury severity.
- traffic safety,
- data-driven,
- assessment,
- random forest,
- vulnerable road user
Citation: Ye Yu, Zhiyuan Liu. A data-driven on-site injury severity assessment model for car-to-electric-bicycle collisions based on positional relationship and random forest[J]. Electronic Research Archive, 2023, 31(6): 3417-3434. doi: 10.3934/era.2023173

Related Papers:

Abstract

Vulnerable road users (VRUs) are usually more susceptible to fatal injuries. Accurate and rapid assessment of VRU injury severity at the accident scene can provide timely support for decision-making in emergency response. However, evaluating VRU injury severity at the accident scene usually requires medical knowledge and medical devices. Few studies have explored the possibility of using on-site positional relationship to assess injury severity, which could provide a new perspective for on-site transportation professionals to assess accident severity. This study proposes a data-driven on-site injury severity assessment model for car-to-electric-bicycle accidents based on the relationship between the final resting positions of the car, electric bicycle and cyclist at the accident scene. Random forest is employed to learn the accident features from the at-scene positional relationship among accident participants, by which injury severity of the cyclist is assessed. Conditional permutation importance, which can account for correlation among predictor variables, is adopted to reflect the importance of predictor variables more accurately. The proposed model is demonstrated using simulated car-to-electric-bicycle collision data. The results show that the proposed model has good performance in terms of overall accuracy and is balanced in recognizing both fatal and non-fatal accidents. Model performance under partial information confirms that the position information of the electric bicycle is more important than the position information of the cyclist in assessing injury severity.

References

[1]	World Health Organization, Global Status Report on Road Safety 2018, 2018. Available from: https://www.who.int/publications/i/item/9789241565684.
[2]	K. Santos, J. P. Dias, C. Amado, A literature review of machine learning algorithms for crash injury severity prediction, J. Saf. Res., 80 (2022). 254–269. https://doi.org/10.1016/j.jsr.2021.12.007 doi: 10.1016/j.jsr.2021.12.007
[3]	J. Park, M. Abdel-Aty, Application of random effects nonlinear model for analyzing motorized and nonmotorized traffic safety performance, J. Transp. Eng. Part A. Syst., 147 (2021), 04020147. https://doi.org/10.1061/jtepbs.0000485 doi: 10.1061/jtepbs.0000485
[4]	J. A. Alagbe, H. H. Han, S. Jin, Effect of technological distractions on pedestrian safe-crossing performance during mixed pedestrian-bicycle flow overlapping with turning vehicles: A case study of Hangzhou, China, J. Transp. Eng. Part A. Syst., 149 (2023), 05022007. https://doi.org/10.1061/jtepbs.Teeng-7597 doi: 10.1061/jtepbs.Teeng-7597
[5]	W. Chen, F. Zhu, Discussions on pedestrian delay models and applications at signalized crosswalks, Multimodal Transp., 1 (2022), 100039. https://doi.org/10.1016/j.multra.2022.100039 doi: 10.1016/j.multra.2022.100039
[6]	H. Ding, N. N. Sze, Effects of road network characteristics on bicycle safety: A multivariate Poisson-lognormal model, Multimodal Transp., 1 (2022), 100020. https://doi.org/10.1016/j.multra.2022.100020 doi: 10.1016/j.multra.2022.100020
[7]	Q. Yuan, X. Xu, T. Wang, Y. Chen, Investigating safety and liability of autonomous vehicles: Bayesian random parameter ordered probit model analysis, J. Intell. Connected Veh., 5 (2022), 199–205. https://doi.org/10.1108/JICV-04-2022-0012 doi: 10.1108/JICV-04-2022-0012
[8]	Y. J. Hu, Y. Zhang, K. S. Shelton, Where are the dangerous intersections for pedestrians and cyclists: A colocation-based approach, Transp. Res. Part C Emerging Technol., 95 (2018), 431–441. https://doi.org/10.1016/j.trc.2018.07.030 doi: 10.1016/j.trc.2018.07.030
[9]	M. G. Mohamed, N. Saunier, L. F. Miranda-Moreno, S. V. Ukkusuri, A clustering regression approach: A comprehensive injury severity analysis of pedestrian-vehicle crashes in New York, US and Montreal, Canada, Saf. Sci., 54 (2013), 27–37. https://doi.org/10.1016/j.ssci.2012.11.001 doi: 10.1016/j.ssci.2012.11.001
[10]	A. Kumar, M. Paul, I. Ghosh, Analysis of pedestrian conflict with right-turning vehicles at signalized intersections in India, J. Transp. Eng. Part A. Syst., 145 (2019), 04019018. https://doi.org/10.1061/jtepbs.0000239 doi: 10.1061/jtepbs.0000239
[11]	R. O. Mujalli, L. Garach, G. Lopez, T. Al-Rousan, Evaluation of injury severity for pedestrian-vehicle crashes in Jordan using extracted rules, J. Transp. Eng. Part A. Syst., 145 (2019), 04019028. https://doi.org/10.1061/jtepbs.0000244 doi: 10.1061/jtepbs.0000244
[12]	I. Isaksson-Hellman, J. Toreki, The effect of speed limit reductions in urban areas on cyclists' injuries in collisions with cars, Traffic Inj. Prev., 20 (2019). S39–S44. https://doi.org/10.1080/15389588.2019.1680836 doi: 10.1080/15389588.2019.1680836
[13]	C. Leo, C. Klug, M. Ohlin, N. M. Bos, R. J. Davidse, A. Linder, Analysis of Swedish and Dutch accident data on cyclist injuries in cyclist-car collisions, Traffic Inj. Prev., 20 (2019). S160–S162. https://doi.org/10.1080/15389588.2019.1679551 doi: 10.1080/15389588.2019.1679551
[14]	A. K. Hoye, O. Johansson, I. S. Hesjevoll, Safety equipment use and crash involvement among cyclists - Behavioral adaptation, precaution or learning?, Transp. Res. Part F Psychol. Behav., 72 (2020), 117–132. https://doi.org/10.1016/j.trf.2020.05.002 doi: 10.1016/j.trf.2020.05.002
[15]	G. Li, Z. Yang, Y. Y. Pan, J. X. Ma, Analysing and modelling of discretionary lane change duration considering driver heterogeneity, Transportmetrica B: Transport Dyn., 11 (2023), 343–360. https://doi.org/10.1080/21680566.2022.2067599. doi: 10.1080/21680566.2022.2067599
[16]	K. Huang, C. Jiang, P. Li, A. Shan, J. Wan, W. H. Qin, A systematic framework for urban smart transportation towards traffic management and parking, Electron. Res. Arch., 30 (2022), 4191–4208. https://doi.org/10.3934/era.2022212 doi: 10.3934/era.2022212
[17]	Y. Liu, R. Jia, J. Ye, X. Qu, How machine learning informs ride-hailing services: A survey, Commun. Transp. Res., 2 (2022), 100075. https://doi.org/10.1016/j.commtr.2022.100075 doi: 10.1016/j.commtr.2022.100075
[18]	S. Li, Y. Liu, X. B. Qu, Model controlled prediction: A reciprocal alternative of model predictive control, IEEE/CAA J. Autom. Sin., 9 (2022), 1107–1110. https://doi.org/10.1109/jas.2022.105611 doi: 10.1109/jas.2022.105611
[19]	Y. Liu, F. Y. Wu, C. Lyu, S. Li, J. P. Ye, X. B. Qu, Deep dispatching: A deep reinforcement learning approach for vehicle dispatching on online ride-hailing platform, Transp. Res. E: Logist. Transp. Rev., 161 (2022). https://doi.org/10.1016/j.tre.2022.102694 doi: 10.1016/j.tre.2022.102694
[20]	Y. Zhang, Q. Cheng, Y. Liu, Z. Liu, Full-scale spatio-temporal traffic flow estimation for city-wide networks: a transfer learning based approach, Transportmetrica B: Transport Dyn., 11 (2022), 869–895. https://doi.org/10.1080/21680566.2022.2143453 doi: 10.1080/21680566.2022.2143453
[21]	Z. Y. Sun, Y. X. Xing, J. Y. Wang, X. Gu, H. P. Lu, Y. Y. Chen, Exploring injury severity of vulnerable road user involved crashes across seasons: A hybrid method integrating random parameter logit model and Bayesian network, Saf. Sci., 150 (2022), 105682. https://doi.org/10.1016/j.ssci.2022.105682 doi: 10.1016/j.ssci.2022.105682
[22]	A. Islam, M. Mekker, P. A. Singleton, Examining pedestrian crash frequency, severity, and safety in numbers using pedestrian exposure from utah traffic signal data, J. Transp. Eng. Part A. Syst., 148 (2022), 04022084. https://doi.org/10.1061/jtepbs.0000737 doi: 10.1061/jtepbs.0000737
[23]	G. Fountas, A. Fonzone, A. Olowosegun, C. McTigue, Addressing unobserved heterogeneity in the analysis of bicycle crash injuries in Scotland: A correlated random parameters ordered probit approach with heterogeneity in means, Anal. Methods Accid. Res., 32 (2021), 100181. https://doi.org/10.1016/j.amar.2021.100181 doi: 10.1016/j.amar.2021.100181
[24]	A. Behnood, S. H. Hosseini, S. R. Davoodi, Bicyclists injury severities: An empirical assessment of temporal stability, Accid. Anal. Prev., 168 (2022), 106616. https://doi.org/10.1016/j.aap.2022.106616 doi: 10.1016/j.aap.2022.106616
[25]	L. Breiman, Random forests, Mach. Learn., 45 (2001), 5–32. https://doi.org/10.1023/a:1010933404324 doi: 10.1023/a:1010933404324
[26]	C. Zhang, J. Bin, W. Wang, X. Peng, R. Wang, R. Halldearn, et al., AIS data driven general vessel destination prediction: A random forest based approach, Transp. Res. Part C Emerging Technol., 118 (2020), 102729. https://doi.org/10.1016/j.trc.2020.102729 doi: 10.1016/j.trc.2020.102729
[27]	B. Dadashova, B. Arenas-Ramires, J. Mira-McWillaims, K. Dixon, D. Lord, Analysis of crash injury severity on two trans-European transport network corridors in Spain using discrete-choice models and random forests, Traffic Inj. Prev., 21 (2020), 228–233. https://doi.org/10.1080/15389588.2020.1733539 doi: 10.1080/15389588.2020.1733539
[28]	H. Bai, L. Li, Y. Wu, C. Liu, Z. Gong, G. Feng, et al., Study on the influence of meteorological elements on growing season vegetation coverage in Xinjiang, China, Electron. Res. Arch., 30 (2022), 3463–3480. https://doi.org/10.3934/era.2022177 doi: 10.3934/era.2022177
[29]	M. X. Xu, H. Y. Lin, Y. Liu, A deep learning approach for vehicle velocity prediction considering the influence factors of multiple lanes, Electron. Res. Arch., 31 (2022), 401–420. https://doi.org/10.3934/era.2023020 doi: 10.3934/era.2023020
[30]	W. R. Gao, Z. H. Bai, F. Zhu, C. C. Chou, B. H. Jiang, A study on the cyclist head kinematic responses in electric-bicycle-to-car accidents using decision-tree model, Accid. Anal. Prev., 160 (2021), 106305. https://doi.org/10.1016/j.aap.2021.106305 doi: 10.1016/j.aap.2021.106305
[31]	Y. Meng, C. Untaroiu, Numerical investigation of occupant injury risks in car-to-end terminal crashes using dummy-based injury criteria and vehicle-based crash severity metrics, Accid. Anal. Prev., 145 (2020), 105700. https://doi.org/10.1016/j.aap.2020.105700 doi: 10.1016/j.aap.2020.105700
[32]	J. Xu, S. Shang, G. Z. Yu, H. S. Qi, Y. P. Wang, S. C. Xu, Are electric self-balancing scooters safe in vehicle crash accidents?, Accid. Anal. Prev., 87 (2016), 102–116. https://doi.org/10.1016/j.aap.2015.10.022 doi: 10.1016/j.aap.2015.10.022
[33]	N. R. Garge, G. Bobashev, B. Eggleston, Random forest methodology for model-based recursive partitioning: the mobForest package for R, BMC Bioinf., 14 (2013), 125. https://doi.org/10.1186/1471-2105-14-125 doi: 10.1186/1471-2105-14-125
[34]	R. Genuer, J. M. Poggi, C. Tuleau-Malot, Variable selection using random forests, Pattern Recognit. Lett., 31 (2010), 2225–2236. https://doi.org/10.1016/j.patrec.2010.03.014 doi: 10.1016/j.patrec.2010.03.014
[35]	C. Strobl, A. L. Boulesteix, T. Kneib, T. Augustin, A. Zeileis, Conditional variable importance for random forests, BMC Bioinf., 9 (2008), 307. https://doi.org/10.1186/1471-2105-9-307 doi: 10.1186/1471-2105-9-307
[36]	K. J. Archer, R. V. Kirnes, Empirical characterization of random forest variable importance measures, Comput. Stat. Data Anal., 52 (2008), 2249–2260. https://doi.org/10.1016/j.csda.2007.08.015 doi: 10.1016/j.csda.2007.08.015
[37]	C. Strobl, A. L. Boulesteix, A. Zeileis, T. Hothorn, Bias in random forest variable importance measures: Illustrations, sources and a solution, BMC Bioinf., 8 (2007), 25. https://doi.org/10.1186/1471-2105-8-25 doi: 10.1186/1471-2105-8-25
[38]	C. Sammut, G. I. Webb, Encyclopedia of Machine Learning and Data Mining, 2^nd edition, Springer, New York, 2017.

Reader Comments

Your name:*

Email:*
© 2023 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)