1 School of Automotive and Transportation Engineering, Hefei University of Technology, Hefei 230009, China; 2 State Key Laboratory of Cognitive Intelligence, Hefei 230009, China
Abstract In order to analyze and learn the difference in car-following behavior between normal and rainy days, we first collect car-following trajectory data of an urban elevated road on normal and rainy days by microwave radar and analyze the differences in speed, relative speed, acceleration, space headway, and time headway among data through statistics. Secondly, owing to the time-series characteristics of car-following data, we use the long short-term memory (LSTM) neural network optimized by attention mechanism (AM) and sparrow search algorithm (SSA) to learn the different car-following behaviors under different weather conditions and build corresponding models (ASL-Normal, ASL-Rain, where ASL stands for AM-SSA-LSTM), respectively. Finally, the simulation test shows that the mean square error (MSE) and reciprocal of time-to-collision (RTTC) of the ASL model are better than those of LSTM and intelligent diver model (IDM), which is closer to the real data. The ASL model can better learn different driving behaviors on normal and rainy days. However, it has a higher sensitivity to weather conditions from cross test on normal and rainy data-sets which need classification training or sample diversification processing. In the car-following platoon simulation, the stability performances of two models are excellent, which can describe the basic characteristics of traffic flow on normal and rainy days. Comparing with ASL-Rain model, the convergence time of ASL-Normal is shorter, reflecting that cautious driving behavior on rainy days will reduce traffic efficiency to a certain extent. However, ASL-Normal model produces a more severe and frequent traffic oscillation within a shorter period because of aggressive driving behavior on normal days.
(Granular models of complex systems; traffic flow)
Fund: Project supported by the National Natural Science Foundation of China (Grant No. 52072108), the Natural Science Foundation of Anhui Province, China (Grant No. 2208085ME148), and the Open Fund for State Key Laboratory of Cognitive Intelligence, China (Grant No. W2022JSKF0504).
Corresponding Authors:
Hai-Jian Bai
E-mail: baihaijian@hfut.edu.cn
Cite this article:
Hai-Jian Bai(柏海舰), Chen-Chen Guo(过晨晨), Heng Ding(丁恒), Li-Yang Wei(卫立阳), Ting Sun(孙婷), and Xing-Yu Chen(陈星宇) Modeling differential car-following behavior under normal and rainy conditions: A memory-based deep learning method with attention mechanism 2023 Chin. Phys. B 32 060507
[1] Hjelkrem O A and Ryeng E O2016 Accid. Anal. Prev.95 227 [2] Chakrabarty N and Gupta K2013 Procedia. Soc. Behav. Sci.104 1048 [3] Soria I, Elefteriadou L and Kondyli A2014 Simul. Modell. Pract. Theory.40 208 [4] Hammit B E, Ghasemzadeh A, James R M, Ahmed M M and Young R K2018 Transp. Res. Part F59 244 [5] Zheng L, Zhu C, He Z B, He T and Liu S S2019 Transportmetrica B7 765787 [6] Li L, Jiang R, He Z B, Chen X Q and Zhou X S2020 Transp. Res. Part C114 225 [7] Treiber M, Hennecke A and Helbing D2000 Phys. Rev. E62 1805 [8] Jiang R, Wu Q S and Zhu Z J2001 Phys. Rev. E64 017101 [9] Gazis D C, Herman R and Rothery R W1961 Oper. Res.9 545 [10] Helbing D and Tilch B1998 Phys. Rev. E58 133 [11] Bando M, Hasebe K, Nakayama A, Shibata A and Sugiyama Y1995 Phys. Rev. E51 1035 [12] Gipps P G1981 Transp. Res. Part B15 105 [13] Newell G F1961 Oper. Res.9 209 [14] Huang Y, Yan X D, Li X M, Duan K, Gao Z J and Rakotonirainy A2022 Transportmetrica A [15] Wei D L and Liu H C2013 Transp. Res. Part B47 1 [16] Papathanasopoulou V and Antoniou C2015 Transp. Res. Part C55 496 [17] He Z B, Zheng L and Guan W2015 Transp. Res. Part B80 185 [18] Yu Y, He Z B and Qu X B2021 IEEE Trans. Cybern.53 1405 [19] Chen T Y, Wong Y D, Shi X P and Yang Y Y 2021 Accid. Anal. Prev.154 106061 [20] Li Y F, Lu X D, Ren C and Zhao H W2019 IEEE Access.7 162778 [21] Kehtarnavaz N, Groswold N, Miller K and Lascoe P1998 IEEE Trans. Veh. Technol.47 694 [22] Panwai S and Dia H F2007 IEEE Trans. Intell. Transp. Syst.8 60 [23] Chong L, Abbas M M and Medina A2011 Transp. Res. Rec.2249 44 [24] Zheng J, Suzuki K and Fujita M2013 Transp. Res. Part C36 339 [25] Zhou M F, Qu X B and Li X P2017 Transp. Res. Part C84 245 [26] Morton J, Wheeler T A and Kochenderfer M J2017 IEEE Trans. Intell. Transp. Syst.18 1289 [27] Huang X L, Sun J and Sun J2018 Transp. Res. Part C95 346 [28] Zhang X H, Sun J, Qi X and Sun J2019 Transp. Res. Part C104 287 [29] Wang X, Jiang R, Li L, Lin Y L and Wang F Y2019 Physica A514 786 [30] Wang X, Jiang R, Li L, Lin Y L, Zheng X H and Wang F Y2018 IEEE Trans. Intell. Transp. Syst.19 910 [31] F. Hui, C. Wei, W. ShangGuan, R. Ando, S. Fang2022 Physica A593 126869 [32] Ma L J and Qu S R2020 Transp. Res. Part C120 102785 [33] Zhu M X, Wang Y H, Pu Z Y, Hu J Y, Wang X S and Ke R M2020 Transp. Res. Part C117 102662 [34] Shi H T, Zhou Y, Wu K S, Wang X, Lin Y X, Ran B2021 Transp. Res. Part C133 103421 [35] Hu Y P, Li Y, Huang H L, Lee J, Yuan C and Zou G Q2022 Accid Anal Prev.165 106503 [36] Ali Y, Sharma A, Haque M M, Zheng Z and Saifuzzaman M2020 Accid Anal Prev.144 105643 [37] Zong F, Wang M, Tang J J and Zeng M2022 Physica A589 126625 [38] Xu J S, Hilker N, Turchet M, Al-Rijleh M K, Tu R, Wang A, Fallahshorshani M, Evans G and Hatzopoulou M2018 Transp. Res. Part D62 90 [39] National Research Council HCM2010: highway capacity manual, 5th edn. 2010 Transp. Res. Board [40] Montanino M and Punzo V2015 Transp. Res. Part B80 82 [41] Khodayari A, Ghaffari A, Kazemi R and Braunstingl R2012 IEEE Trans. Syst. Man. Cybern. Part A42 1440 [42] Thiemann C, Treiber M and Kesting A2008 Transp. Res. Rec: J. Transp. Res. Board.2088 90 [43] Wang Z 2020 Expressway Traffic Flow Characteristic Analysis and Prediction in Rainy Environment, MS dissertation (Beijing: Beijing Jiaotong University) (in Chinese) [44] Xu F J 2017 Calculation and Analysis of Urban Road Capacity Under Rainfall Condition, MS dissertation (Xi'an: Changan University) (in Chinese) [45] Xue J K and Shen B2020 Syst. Sci. Control. Eng.8 2234 [46] Jiang T, Sun H, Dai Y G and Liu D2020 J. Phys. Conf. Ser.1607 012001 [47] Heinrich K, Zschech P, Janiesch C and Bonin M2021 Decis. Support. Syst.143 113494 [48] Zhu M X, Wang X S, Tarko A and Fang S E2018 Transp. Res. Part C93 425 [49] Sangster J, Rakha H and Du J2013 Transp. Res. Rec.2390 20 [50] Qu X B, Wang S A and Zhang J2015 Transp. Res. Part B73 91 [51] Vogel K2003 Accid. Anal. Prev.35 427 [52] Jian M Y and Shi J2020 Saf. Sci.122 104536 [53] Ding H, Pan H, Bai H J, Zheng X Y, Chen J and Zhang W H2022 Physica A596 127154 [54] Chen D J, J. Laval, Zheng Z D and Ahn S2012 Transp. Res. Part B46 744 [55] Laval J A2011 Transp. Res. Part B45 385
[1]
Disruption prediction based on fusion feature extractor on J-TEXT Wei Zheng(郑玮), Fengming Xue(薛凤鸣), Zhongyong Chen(陈忠勇), Chengshuo Shen(沈呈硕), Xinkun Ai(艾鑫坤), Yu Zhong(钟昱), Nengchao Wang(王能超), Ming Zhang(张明),Yonghua Ding(丁永华), Zhipeng Chen(陈志鹏), Zhoujun Yang(杨州军), and Yuan Pan(潘垣). Chin. Phys. B, 2023, 32(7): 075203.
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.
